The Physical Geography of the Mediterranean (Oxford Regional Environments)

The Physical Geography of the Mediterranean

THE OXFORD REGIONAL ENVIRONMENTS SERIES PUBLISHED

The Physical Geography of Africa edited by William M. Adams, Andrew S. Goudie, and Antony R. Orme The Physical Geography of North America edited by Antony R. Orme The Physical Geography of Northern Eurasia edited by Maria Shahgedanova The Physical Geography of Southeast Asia edited by Avijit Gupta The Physical Geography of Fennoscandia edited by Matti Seppälä The Physical Geography of Western Europe edited by Eduard A. Koster The Physical Geography of South America edited by Thomas T. Veblen, Kenneth R. Young, and Antony R. Orme FORTHCOMING

The Physical Geography of the British Isles edited by Adrian Parker

The Physical Geography of the Mediterranean edited by Jamie Woodward

1

3

Great Clarendon Street, Oxford OX2 6DP Oxford University Press is a department of the University of Oxford. It furthers the University’s objective of excellence in research, scholarship, and education by publishing worldwide in Oxford New York Auckland Cape Town Dar es Salaam Hong Kong Karachi Kuala Lumpur Madrid Melbourne Mexico City Nairobi New Delhi Shanghai Taipei Toronto With offices in Argentina Austria Brazil Chile Czech Republic France Greece Guatemala Hungary Italy Japan Poland Portugal Singapore South Korea Switzerland Thailand Turkey Ukraine Vietnam Oxford is a registered trademark of Oxford University Press in the UK and in certain other countries Published in the United States by Oxford University Press Inc., New York © The several contributors 2009 The moral rights of the author have been asserted Database right Oxford University Press (maker) First published 2009 All rights reserved. No part of this publication may be reproduced, stored in a retrieval system, or transmitted, in any form or by any means, without the prior permission in writing of Oxford University Press, or as expressly permitted by law, or under terms agreed with the appropriate reprographics rights organization. Enquiries concerning reproduction outside the scope of the above should be sent to the Rights Department, Oxford University Press, at the address above You must not circulate this book in any other binding or cover and you must impose the same condition on any acquirer British Library Cataloguing in Publication Data Data available Library of Congress Cataloging in Publication Data The physical geography of the Mediterranean / edited by Jamie Woodward. p. cm. – (Oxford regional environments series) ISBN 978–0–19–926803–0 1. Physical geography–Mediterranean Region. 2. Mediterranean Region–Geography. I. Woodward, Jamie C. GB178.P48 2009 508.3182’2–dc22 2009005674 Typeset by SPI Publisher Services, Pondicherry, India Printed in Great Britain on acid-free paper by CPI Antony Rowe, Chippenham, Wiltshire ISBN 978–0–19–926803–0 1 3 5 7 9 10 8 6 4 2

Frontispiece. A reconstruction of ancient Sparta on the alluvial plain of the Evrotas River in southern Greece. The view looks to the west to the peaks of Mount Taygetos. Reproduced with permission from an illustration by J. P. Mahaffy published in 1890.

‘. . . the Mediterranean can’t be reduced down to one landscape or one lifestyle. Could there be a more tenuous link than that which bonds, upon its shores, the luxuriant landscapes of the coast and the arid deserts inland? . . . The Mediterranean lover can recognise the imperceptible variations that alter the texture of a valley, the hue of a city, and the quite special light of a particular bay. But our ideas about the Mediterranean are stubborn. Its diversity doesn’t stop us from seeing a Mediterranean nature, climate and landscape at work all over.’ (Girard, 2001, p. 36) Girard, X. (2001), Mediterranean: from Homer to Picasso. Assouline, New York.

For Sam and Alex

Foreword The Physical Geography of the Mediterranean is the eighth in a series of advanced books that is being published by Oxford University Press under the rubric of Oxford Regional Environments. The aim of the series is to provide a durable statement of physical conditions on each of the continents, or major regions within those continents. Each volume includes a discussion of the systematic framework of the region (for instance, tectonism, climate, biogeography), followed by an evaluation of dominant environments (such as mountains, forests, and deserts) and their linkages, and concludes with a consideration of the main environmental issues related to the human use and misuse of the land (such as resource exploitation, agricultural and urban impacts, pollution, and nature conservation). While books in the series are framed within an agreed context, individual books seek to emphasize the distinctive qualities of each region. We hope that this approach will provide a coherent and informative basis for physical geography and related sciences, and that each volume will be an important and useful reference source for those concerned with understanding the varied environments of the continents. Andrew Goudie, University of Oxford Antony Orme, University of California, Los Angeles

This page intentionally left blank

Preface and Acknowledgements Scholars have been fascinated by the Mediterranean environment since classical times. The Mediterranean, including the sea itself and the climates, landscapes, and ecosystems around it, forms one of the most intensively studied natural laboratories on Earth and recent years have seen remarkable advances in our understanding of many components of its physical geography. These advances have come from the efforts of countless researchers in many disciplines. This book explores the evolution and functioning of the Mediterranean environment by drawing upon findings derived from studies of modern processes as well as data from long records of change. It assesses the main drivers of environmental change (including tectonic processes and climate change) and their effects— with particular emphasis on the glacial-interglacial cycles of the Quaternary ice age. Several chapters also draw upon sedimentary, archaeological and historical records to examine the nature of human-environment interactions during the postglacial period in order to explore the role of humans in shaping the landscapes and habitats we see today. This book examines these processes and the key debates, and places natural hazards and current environmental issues in long term context. The physical geography of the Mediterranean involves a big canvas and a vast literature. This book is the first modern synthesis of the physical geography of the entire Mediterranean region and it was conceived, from the outset, as a multi-authored volume and as an international, multidisciplinary team effort. It incorporates the talents and experience of thirty five scholars who, between them, have direct field experience in all of the countries that border the Mediterranean Sea and in all of the major islands and marine basins within it. Many of the contributors are the leaders in their fields. This body of expertise gives this volume an authority that could not be attained by a singleauthored text. The twenty-three chapters are organised into four main parts—each with an Editorial Introduction that draws out major themes—under the following headings: I. II. III. IV.

The Physical and Biological Framework (Chapters 1 to 5) Process and Change in Specific Environments (Chapters 6 to 14) Hazards (Chapters 15 to 19) Environmental Issues in the 21st Century (Chapters 20 to 23)

This book will appeal to all scholars and students of geography, Earth science and ecology who are involved in the study of the Mediterranean and all who are interested, more broadly, in Mediterranean environments, geomorphology, natural hazards, Quaternary environmental change, biodiversity and conservation, and the human impact on the natural environment. The last decade has seen the publication of several excellent books on the history and prehistory of the Mediterranean world and something of a revival in considering the entire region as a unit of study. Research into the human past in the Mediterranean cannot be separated from an understanding of the opportunities and constraints offered by such a dynamic (and often hazardous) and varied environment. I hope, therefore, that this book will also have wide appeal to archaeologists and historians of the Mediterranean world and to all who are interested, more generally, in humanenvironment interactions—especially over extended timescales Interacting with so many authors and disciplines has been a challenging but hugely rewarding experience and I would like to offer my warmest thanks to all the authors for their contributions and for their patience and cooperation during the reviewing process and the final editorial and production stages. Before final editing, all

x

Preface and Acknowledgements

twenty-three chapters were formally reviewed by the editor and by at least one external reviewer and I would like to thank the following for their generous help in this task: Clive Agnew (Manchester), Harriet Allen (Cambridge), Nick Ambraseys (London), Grant Bigg (Sheffield), Louise Bracken (Durham), Rob Bryant (Sheffield), David Chester (Liverpool), Jacques-Louis de Beaulieu (Marseilles), Simon Davis (Lisbon), Mick Frogley (Sussex), Dick Grove (Cambridge), Philip Hughes (Manchester), Ian Lawson (Leeds), Mike Leeder (Norwich), John Lewin (Aberystwyth), Sarah Lindley (Manchester), Donatella Magri (Milan), Allen Perry (Swansea), Rick Shakesby (Swansea), Heather Viles (Oxford), Des Walling (Exeter), Tony Waltham (Nottingham), and Kathy Willis (Oxford). Jeff Blackford, Karen Exell, Philip Hughes, John Lewin, and Chris Perkins also provided helpful comments on the four Editorial Introductions. In the early stages of this project the series editors, Andrew Goudie and Antony Orme, provided very helpful feedback on the initial book proposal. I am also grateful for the support of my colleagues at the University of Manchester and to Tim Allott (Head of Geography) who provided funding to cover the costs of translating Chapter 21. Avi Gupta provided wise words of support—especially in the early stages when his volume, The Physical Geography of Southeast Asia, was nearing completion. I am also very grateful to Anne Ashby, Dominic Byatt, Lizzy Suffling, and Louise Sprake at Oxford University Press who have provided valuable advice and encouragement at various stages. I would also like to thank Sylvie Jaffrey and Debbie Sutcliffe for all their help at the copy-editing and proof checking stages. Of the many people who have helped along the way getting this book to publication, I would especially like to thank Nick Scarle from the Cartographic Unit at the University of Manchester. Nick managed the figure and photograph database for this book and redrew most of the maps and figures from scratch. This was a huge undertaking, but as the figures and photographs arrived in Manchester from distant lands via email, CD, and hard copy (at least two were barely decipherable scans of sketches made on paper napkins) Nick prepared them for publication with his trademark expertise and endless patience and good humour. I would also like to thank Graham Bowden in the Cartographic Unit at Manchester who worked on the figures in Chapters 13 and 17. I would also like to thank John Prag at the Manchester Museum who provided the source book for the frontispiece. My own research in the Mediterranean began in 1986 during my Ph.D. research in Cambridge on the Klithi Project in north-west Greece directed by Geoff Bailey (Bailey, 1997). I was fortunate at that time to cut my teeth in the field with Mark Macklin and John Lewin and our collaborations have continued ever since. I am delighted that they were able to co-author chapters with me in the current volume. I must also thank all the undergraduate students at Leeds and Manchester who either took one of the various incarnations of my final year course on Mediterranean Quaternary Environments or participated in the field course to south-east Spain—their comments and feedback have helped to shape this book. I am also extremely grateful to the following research students who have participated in Mediterranean adventures: George Christopolos, Graham Smith, Rob Hamlin, Suzanne Hewitt, Maroulia Zorzou, Philip Hughes, Mike Morley, and Rose Wilkinson. My research in the Mediterranean has involved collaboration with many projects and individuals and I have been fortunate to receive support and encouragement from many colleagues. I would especially like to thank the following: Geoff Bailey, Graeme Barker, Alex Chepstow-Lusty, Ian Foster, Mick Frogley, Clive Gamble, Philip Gibbard, David Gilbertson, Paul Goldberg, Dick Grove, Philip Hughes, Takis Karkanas, Mike Kirkby, Eleni Kotjabopoulou, Mike Krom, Henry Lamb, John Lewin, Mark Macklin, Rolfe Mandel, Mark Pluciennik, Jim Rose, Nick Shackleton, John Thornes, Charles Turner, Chronis Tzedakis, Claudio Vita-Finzi, Richard West, Bob Whallon, and Martin Williams. I would also like to pay a special tribute to John Thornes who died in the summer of 2008 when this book was in the final stages of production. John was an inspirational

Preface and Acknowledgements

xi

figure in physical geography who was passionate about the Mediterranean, its people, and its geomorphology. For many years John coordinated and led the MEDALUS projects funded by the European Union and his research teams produced many key insights into land degradation and river basin processes in the semi-arid Mediterranean. John wrote two chapters for this book (one of them co-authored with me and Francisco LópezBermúdez) and it was a great pleasure to be able to work closely with him throughout. He was a stimulating, supportive, and generous collaborator and I hope that he would have been pleased with this book. Finally I must thank Jenny, Sam, and Alex for their love and support. J. W. Manchester November 2008

Reference Bailey, G. N. (ed.) (1997), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece (2 vols.) McDonald Institute for Archaeological Research, Cambridge. The authors, editor, and publisher thank the following who have kindly given permission for the modification and use of copyright material: Frontispiece: ‘Ancient Sparta’ in Greek Pictures, p. 173, J.P. Mahaffy (1890) published by the Religious Tract Society. Reproduced with permission from Lutterworth Press. Fig. 1.7: from Junta de Andalucía Ortofotografía digital de Andalucía (ISBN 84-95083-96-5) with permission. Figs. 3.3(a) and (b): from Futures for the Mediterranean Basin: The Blue Plan, 6–7, Grenon and Batisse (1989) with permission from Oxford University Press. Table 3.4: from http://natural-hazards.jrc.ec.europa.eu/activities_flood_flashflood.html with permission from Jutta Thielen-del Pozo at JRC European Commission. Figs. 5.1, 5.3, 5.6, 5.11, 5.12, 5.15, 5.17, 5.20 and Table 5.1: from Biology and Wildlife of the Mediterranean Region, Blondel and Aronson (eds.) (1999) with permission from Oxford University Press. Fig. 6.3: from Catena 40, 3–17, ‘The effect of land parameters on vegetation performance and degree of erosion under Mediterranean conditions’, Kosmas et al. (2000) with permission from Elsevier. Fig. 6.4: from Catena, 28, 157–169, ‘Soils in the Mediterranean region: what makes them different?’ Yaalon (1997) with permission from Elsevier. Fig. 6.9: from Soil and Tillage Research 85, 123–142, Assessment of tillage erosion by mouldboard plough in Tuscany (Italy)’, De Alba et al. (2006) with permission from Elsevier. Fig. 6.11: from Geomorphology 13, 87–99, ‘Short and long term effects of bioturbation on soil erosion, water resources and soil development in an arid environment’, Yair (1995) with permission from Elsevier. Fig. 6.12: from ‘Mechanisms of overland flow generation and sediment production on loamy and sandy soils with and without rock fragments’, Poesen et al. in Overland Flow Hydraulics and Erosion Mechanics, Parsons and Abrahams (eds.) (1992) with kind permission from Professor Jean Poesen. Fig. 6.13: Reproduced with kind permission from Juan Puigdefábregas. Fig. 6.16: from Geomorphology 26, 239–251, ‘Factors underlying piping in the Basilicata region, southern Italy’, Farifteh and Soeters (1999) with permission from Elsevier. Fig. 6.18: from Geoderma 105, 125–140, ‘Soil erosion caused by extreme rainfall events: mapping and quantification in agricultural plots from very detailed digital elevation models’, Martínez Casasnovas et al. (2002) with permission from Elsevier. Fig 6.20: Earth and Planetary Science Letters 195, 169–183, ‘Power-law correlations of landslide areas in central Italy’, Guzzetti et al. (2002) with kind permission from Fausto Guzzetti and Elsevier.

xii

Preface and Acknowledgements Fig. 6.22: from Engineering Geology 70, 109–130, ‘Instability conditions of marly hillslopes: towards landsliding or gullying? The case of the Barcelonnette Basin, south east France’, Maquaire et al. (2003) with permission from Elsevier. Fig. 6.23: from Engineering Geology 58, 89–107, ‘Landslide fatalities and the evaluation of landslide risk in Italy’, Guzzetti (2000) with permission from Elsevier. Fig. 6.24: from Bulletin of Engineering Geology and the Environment 59, 87–97, ‘History of the 1963 Vaiont slide: the importance of geological factors’, Semenza and Ghirotti (2000) with permission from Elsevier. Fig 7.2: from Israel Journal of Botany 39, 481–508, ‘Global change: vegetation, ecosystems, and land use in the southern Mediterranean basin by the mid 21st century’, Le Houerou (1990) with permission from LPP Ltd. Fig 7.5: from ‘The study of plant groupings in the countries surrounding the Mediterranean: some methodological aspects’, Quézel, 87–93, in Ecosystems of the World, vol. 11 Mediterranean-type Shrublands, di Castri et al. (eds.) (1981) with permission from Elsevier. Fig 7.6: from materials in: (1) the Atlas d’Aréologie Périméditerranéenne Daget, (1980) (Maison d’edition: Institut de Botanique, Montpelier) with permission from the author; (2) ‘Definition of the Mediterranean region and the origin of its flora’, Quézel, 8–24, in Plant Conservation in the Mediterranean Area, in Gómez-Campo (ed.) (1985) with permission from Springer (Kluwer); (3) The Holocene 1, 157–161, ‘The recent distribution of Pinus brutia: a reassessment based on dendroarchaeological and dendrohistorical evidence from Israel’ Biger and Liphschitz (1991) with permission from Hodder Headline. Fig 7.13: from The Holocene, p.190 Roberts (1998) with permission from Blackwell. Fig 8.2: from Futures for the Mediterranean Basin: The Blue Plan, p. 28, Grenon and Batisse (1989) with permission from Oxford University Press. Fig 8.3: from Futures for the Mediterranean Basin: The Blue Plan, p. 221, Grenon and Batisse (1989) with permission from Oxford University Press. Fig. 8.8: from Geoderma, 15, 61–70, ‘Vegetal cover to estimate soil erosion hazard in Rhodesia’, Elwell and Stocking (1976) with permission from Elsevier Science Table 8.1 from Journal of the Geological Society, London 162, 879–908, ‘Recent evolution of a Mediterranean deltaic coastal zone: human impacts on the Inner Thermaikos Gulf’, Kapsimalis et al. (2005) with permission from The Geological Society of London. Fig. 8.16: from ‘Erosion and sediment yield in mountain areas of the world’ Dedkov and Moszherin, 29–36, in Erosion, Debris Flows and Environment in Mountain Regions of the World, Walling et al. (eds.) IAHS Publication No. 209 (1992) with permission from The International Association of Hydrological Sciences Press. Fig. 10.12: from Catena Supplement 25, p. 88, ‘Environmental change and human impacts on the Mediterranean karsts of France, Italy and the Dinaric region’, Gams et al. (1993) with permission from Catena Verlag. Fig. 10.13(a) and (c): from Zeitschrift für Geomorphologie 22, 170–81, ‘The Polje: The problems of its definition’, Gams (1978) with permission from Gebrüder Borntraeger Science Publishers (BerlinStuttgart). Figs. 10.15, 10.16(a) and (c): with permission from Dr Tony Waltham. Fig 10.20: from Catena Supplement 25, p. 8, ‘Environmental change and human impact on karst terrains’, Williams (1993) with permission from Catena Verlag. Fig 11.2(a): from Proceedings of the Prehistoric Society, 32, 1-29, ‘The climate, environment and industries of Stone Age Greece: Part II’, Higgs and Vita-Finzi (1966) with permission from The Prehistoric Society. Fig 11.2(b): from The Mediterranean Valleys: Geological Changes in Historical Times, p. 92, Vita-Finzi (1969), Cambridge, Cambridge University Press, with permission from the author. Fig 11.4(b): from The Mediterranean Valleys: Geological Changes in Historical Times, p. 10, Vita-Finzi (1969), Cambridge, Cambridge University Press, with permission from the author.

Preface and Acknowledgements

xiii

Fig 11.13: from Journal of Field Archaeology, 17, 379–96, ‘Land use and soil erosion in Prehistoric and Historical Greece’, van Andel et al. (1990). Reproduced from Journal of Field Archaeology with permission of the Trustees of Boston University. All rights reserved. Fig 11.14: from ‘Palaeohydrological changes in the Mediterranean region during the Late Quaternary’, Benito p. 131, in Palaeohydology: Understanding Global Change, Gregory and Benito (eds.) (2003). Reproduced with permission of John Wiley & Sons Limited. Fig. 11.16: from Catena 66, 145–54, ‘Past hydrological events reflected in the Holocene fluvial record of Europe’, Macklin et al. (2006) with permission from Elsevier. Fig 11.20: from Journal of the Geological Society, London 162, 879–908, ‘Recent evolution of a Mediterranean deltaic coastal zone: human impacts on the Inner Thermaikos Gulf’, Kapsimalis et al. (2005) with permission from The Geological Society of London. Table 12.2: from Episodes: Journal of International Geoscience 28, 85–92, ‘A formal stratigraphical approach for Quaternary glacial records in mountain regions’, Hughes et al. (2005) with permission from the International Union of Geological Sciences. Fig. 12.8: from Zeitschrift für Gletscherkunde und Glazialgeologie 31, 199–206, ‘Little Ice Age glacier fluctuations in the Pyrenees’, Grove and Gellatly (1995) with permission from the publishers: Universitätsverlag Wagner, Innsbruck. Fig. 12.9: from Little Ice Ages: Ancient and Modern (Volume 1), p. 190, Grove (2004), London, Routledge with permission from the publishers from Cengage Learning Services Limited. Fig. 12.19: from Global and Planetary Change, 50, p. 94, ‘Late Pleistocene glaciers and climate in the Mediterranean’, Hughes et al. (2006) with permission from Elsevier. Table 13.1: from Futures for the Mediterranean Basin: The Blue Plan, p. 31, Grenon and Batisse (1989) with permission from Oxford University Press. Fig. 13.1(a) and (b): maps designed by Dr Thomas Dewez based on the following datasets for topography (STRM30), bathymetry (ETOPO2), and earthquake locations (NEIC). Reproduced with permission from Dr Thomas Dewez (BRGM), French Geological Survey, Orléans, France. Fig. 13.3: from ‘Littoral cells’, Inman 594-599, in Encyclopedia of Coastal Science Schwartz (ed.) (2005). Reproduced with kind permission of Springer Science and Business Media and based on an original figure in: Proceedings of the 19th Coastal Engineering Conference, American Society of Civil Engineers, 2, 1600–17, ‘The Nile littoral cell and man’s impact on the coastal zone of the southeastern Mediterranean’, Inman and Jenkins (1984). Fig. 13.5: from Quaternary Science Reviews, 24, 1969–88, ‘Sea-level change in the Mediterranean since the LGM: model predictions for tectonically stable areas’, Lambeck and Purcell (2005) with permission from Elsevier. Fig. 13.6: from Puglia 2003—Final Conference Project IGCP 437, Sea level change at Capo Caccia (Sardinia) and Mallorca (Balearic Islands) during oxygen isotope sub-stage 5e, based on Th/U datings of phreatic overgrowths on speleothems’, Tuccimei et al. (2003) and originally published in Geodinamica Acta 15, 113–25, ‘Phreatic overgrowths on speleothems: a useful tool in structural geology in littoral karstic landscapes. The example of eastern Mallorca (Balearic Islands)’, Fornós et al. (2002) with permission from Elsevier. Fig. 13.7: from Marine Geology 167, 105–26, ‘Holocene tectonic uplift patterns in northeastern Sicily: evidence from marine notches in coastal outcrops’, Rust and Kershaw (2000) with permission from Elsevier. Fig. 15.12: Reproduced by kind permission of the Syndics of Cambridge University Library. Figs.18.9(a) and (b): Reproduced by kind permission of the Centre Méditerranéen de l’Environment (http://www.cme-cpie84.org/) Fig. 18.14: Reproduced by kind permission of the Conselleria de Medi Ambient del Govern de les Illes Balears. Figs. 19.4 and 19.8: reproduced from open access data from the European Forest Fire Information System (EFFIS) at the European Institute for Environment and Sustainability, EC Joint Research Centre:

xiv

Preface and Acknowledgements Figs. 20.1, 20.5 and 20.10: from Panarchy: Understanding Transformations in Human and Natural Systems, Gunderson and Holling (eds.) (2002) with permission from Island Press, Washington. Fig. 20.4: from Geoderma, 15, 61–70, ‘Vegetal cover to estimate soil erosion hazard in Rhodesia’, Elwell and Stocking (1976) with permission from Elsevier Science Fig. 20.6(b): from ‘The problem’ Kirkby, 1–16, in Soil Erosion, Kirkby and Morgan (eds.) (1980). Reproduced with permission of John Wiley & Sons Limited. Fig. 20.6(c): from ‘The effect of land use on soil erosion and land degradation under Mediterranean conditions’, Kosmas et al. 57–81, in Mediterranean Desertification: A Mosaic of Processes and Responses, Geeson et al. (eds.). Reproduced with permission of John Wiley & Sons Limited. Fig. 20.9: from ‘Erosion-vegetation competition in a stochastic environment undergoing climatic change’, Thornes and Brandt p. 314, in Environmental Change in Drylands: Biogeographical and Geomorphological Perspectives, Millington and Pye (eds.) (1994). Reproduced with permission of John Wiley & Sons Limited. Fig. 23.6: from Biology and Wildlife of the Mediterranean Region edited by Blondel and Aronson (1999) with permission from Oxford University Press. Full references for these figures and tables have been included in their respective chapters. Credits for photographs are given in the respective captions. While every reasonable effort has been made to trace and contact copyright holders, this has not always been successful. We apologize for any apparent negligence. If this list contains errors or inconsistencies, please contact the editor so that these can be corrected in any future editions.

Contents List of Figures List of Tables List of Contributors

xvii xxx xxxii I. The Physical and Biological Framework

Editorial Introduction

3

JAMIE WOODWARD

1. Tectonic Setting and Landscape Development

5

ANNE MATHER

2. The Marine Environment: Present and Past EELCO ROHLING , RAMADAN ABU - ZIED , JAMES CASFORD , ANGELA HAYES ,

33

AND BABETTE HOOGAKKER

3. The Climate System ANDREW HARDING , JEAN PALUTIKOF , AND TOM HOLT

69

4. Cenozoic Climate and Vegetation Change

89

CHRONIS TZEDAKIS

5. The Nature and Origin of the Vertebrate Fauna

139

JACQUES BLONDEL

II. Process and Change in Specific Environments Editorial Introduction

167

JAMIE WOODWARD

6. Weathering, Soils, and Slope Processes

169

JOHN WAINWRIGHT

7. Vegetation and Ecosystem Dynamics

203

HARRIET ALLEN

8. Hydrology, River Regimes, and Sediment Yield JOHN THORNES , FRANCISCO LÓPEZ - BERMÚDEZ , AND JAMIE WOODWARD

229

9. Lakes, Wetlands, and Holocene Environmental Change

255

NEIL ROBERTS AND JANE REED

10. Karst Geomorphology and Environmental Change

287

JOHN LEWIN AND JAMIE WOODWARD

11. River Systems and Environmental Change

319

MARK MACKLIN AND JAMIE WOODWARD

12. Glacial and Periglacial Environments

353

PHILIP HUGHES AND JAMIE WOODWARD

13. Coastal Geomorphology and Sea-Level Change

385

IAIN STEWART AND CHRISTOPHE MORHANGE

14. Aeolian Processes and Landforms ANDREW GOUDIE

415

xvi

Contents

III. Hazards Editorial Introduction

433

JAMIE WOODWARD

15. Volcanoes

435

CLIVE OPPENHEIMER AND DAVID PYLE

16. Earthquakes

469

STATHIS STIROS

17. Tsunamis

493

GERASSIMOS PAPADOPOULOS

18. Storms and Floods

513

MARÍA DEL CARMEN LLASAT

19. Wildfires

541

FRANCISCO LLORET , JOSEP PIÑOL , AND MARC CASTELLNOU

IV. Environmental Issues in the 21st Century Editorial Introduction

561

JAMIE WOODWARD

20. Land Degradation

563

JOHN THORNES

21. Water Resources

583

JEAN MARGAT

22. Air Pollution and Climate

599

JOS LELIEVELD

23. Biodiversity and Conservation

615

JACQUES BLONDEL AND FRÉDÉRIC MÉDAIL

Index

651

List of Figures 1.1. (a) The Alpine Himalayan orogen in its global setting and (b) the main tectonic landform features of the Mediterranean 1.2. Simplified cross-section of subduction rollback 1.3. The present geodynamic framework of the Mediterranean 1.4. Seismic activity in the Mediterranean 1.5. Volcanic activity in the western Mediterranean over the last 33 Ma 1.6. Eastward migration of the topography in conjunction with the eastward rollback of the subduction zone in the Apennines of Italy through time 1.7. An example of alluvial fans from a faulted mountain front in the Tabernas basin of south-east Spain 1.8. An example of well-developed badlands in Tortonian marls within the Tabernas basin of south-east Spain 1.9. Volcanic activity and tsunami impact in the Aeolian Islands 1.10. Plate boundaries and the main poles of rotation 1.11. Oblique view of the Megara basin 1.12. Distribution of erosional and landslide features in relation to the 70 ka river capture site in the Río de Aguas basin, south-east Spain 1.13. Image of an active mass failure along the margins of the Río de Aguas valley 1.14. Schematic evolution of the valley systems of the Sorbas basin before and after the 70 ka river capture 1.15. Surface lowering above the 70 ka river capture site depicted in Figure 1.12 2.1. Map of the Mediterranean Sea 2.2. Longitudinal cross-section showing water mass circulation in the Mediterranean Sea during the present-day winter 2.3. Surface water circulation in the Mediterranean Sea 2.4. Northern Hemisphere summer atmospheric circulation pattern 2.5. Schematic illustration of surface circulation in the Alboran Sea 2.6. Schematic illustration of the main gyres associated with Atlantic surface flow 2.7. Typical salinity profiles for the western and eastern Mediterranean basins 2.8. Schematic illustration of the preconditioning, violent mixing, and deep convection phases 2.9. Estimated sea surface salinity distribution for (a) the Holocene Climate Optimum and (b) the Last Glacial Maximum 2.10. Annual sea surface temperature (SST) reconstructions for the Holocene Climate Optimum and Last Glacial Maximum 2.11. Scanning electron microscope images of the carbonate shells of several planktonic foraminiferal species that live in the Mediterranean Sea 2.12. Example of a laminated sapropel in a freshly opened sediment core 2.13. Schematic presentation of the changes in subsurface circulation patterns between the present day and times of sapropel formation

6 8 9 12 13

17 18 19 20 21 22 24 25 25 26 34 34 35 38 41 41 42 43 48 49 50 51 52

xviii

List of Figures

2.14. Phase relationships between the sapropel record and associated ‰18 O record from core RC9-181 and the orbital cycles of precession and eccentricity 3.1. The location of the Mediterranean region in relation to the large-scale atmospheric circulation 3.2. Seasonal temperature and rainfall variations at selected sites around the Mediterranean 3.3. (a) Mean annual rainfall and (b) length of the dry season across the Mediterranean basin 3.4. (a) Regions of cyclone genesis and dominant cyclone tracks in the Mediterranean and (b) a TERRA satellite image of a cyclone centred on the Ionian Sea 3.5. Composite graphs of annual, winter, and summer temperature in the whole Mediterranean and the western, central, and eastern basins, 1960–2000 3.6. Composite graphs of annual, winter, and summer precipitation in the whole Mediterranean and the western, central, and eastern basins,1960–2000 3.7. Future changes in temperature and precipitation over the Mediterranean 3.8. Future changes in (a) length of the summer drought and (b) maximum five-day precipitation over the Mediterranean 4.1. Location of sites discussed in Chapter 4 4.2. Global oxygen isotope record based on data from more than forty DSDP and ODP sites 4.3. A section of the compiled oxygen isotope record of Figure 4.2 for the interval 4–7 Ma 4.4. Northern Hemisphere palaeogeography and global vegetation maps for selected time slices in the Tertiary 4.5. Summary pollen diagram showing the main taxa of the interglacial succession during the Last Interglacial at Ioannina 4.6. (a) Variations in ‰18 O composition of benthic foraminifera in V19-30 in the East Pacific. (b) Variations in ‰18 O composition of ice in Greenland Ice Sheet Project 2 record. (c) Variations in alkenone-derived sea surface temperatures in marine core MD95-2043 from the Alboran Sea, western Mediterranean. (d) Interval of maximum lake levels of Lake Lisan, Dead Sea Transform area. (e) Interval of Kastritsa beach deposits, Ioannina basin, north-west Greece. (f) Temperate tree pollen percentages curves from Ioannina 1—284, Kopais k93, central Greece 4.7. June insolation for 65˚ N and variations in ‰18 O composition of benthic foraminifera over the last 3 Myr in the Shackleton 06 (S06) composite record from sites in the equatorial East Pacific 4.8. SPOT imagery of the Ioannina basin and surrounding areas showing the extent of topographical variability in the region 4.9. Location of some Mediterranean pollen records from wetland sites spanning all or part of the LGM, with inferred refugial tree populations 4.10. ‘Serial extinction’ of a number of genera in Europe 4.11. Variations in ‰18 O composition of benthic foraminifera in the S06 composite record and arboreal pollen percentages at Tenaghi Philippon, north-eastern Greece

54 70 71 72

76

81

82 84 85 90 92 95 96 101

105

106 108 110 111

114

List of Figures

4.12. Variations in ‰18 O composition of benthic foraminifera in the S06 composite record over the interval 960–1,340 ka 4.13. Variations in ‰18 O composition of ice in Greenland Ice Sheet Project 2 record, and of planktonic foraminifera in core MD95-2042 from the Portuguese margin; variations in alkenone-derived sea surface temperatures in western Mediterranean; temperate tree pollen percentages in marine core MD95-2043 from the Alboran Sea, Ioannina I-284, Kopais K93, and Tenaghi Philippon TF II, Greece 5.1. The four quadrants of the Mediterranean region 5.2. The Lesser kestrel Falco naumanni, a typical but declining species found in old cities and craggy areas in the Mediterranean 5.3. Phylogeography of the Brown bear 5.4. The Brown bear Ursus arctos 5.5. The Ibex Capra ibex 5.6. Patterns of invasion of the Mediterranean basin and western Europe by the house mouse 5.7. Biogeographical origin of the bird fauna of the Mediterranean region 5.8. The Fan-tailed warbler Cisticola juncidis 5.9. Four species of warbler that are typical of Mediterranean matorrals where they evolved: the Mediterranean warbler Sylvia melanocephala; the Subalpine warbler S. cantillans; the Dartford warbler S. undata; and Marmora’s warbler S. sarda 5.10. The Rock partridge Alectoris graeca 5.11. Relationships between the three main groups of Mediterranean warblers (genus Sylvia) and their geographical range 5.12. The zonation of the various vegetation belts in the western Mediterranean area in relation to both altitude and latitude 5.13. Examples of nest types for terns and waders that can be found on small islets within Mediterranean lagoons between April and July 5.14. The Blue tit Cyanistes caeruleus 5.15. The Mediterranean as a key place for migratory and wintering birds 5.16. The Short-toed eagle Circaetus gallicus 5.17. Pygmy hippos and elephants 5.18. Turnover of non-volant mammal species in Corsica as a result of human colonization around 9,500 years ago 5.19. The very large and beautiful Eyed lizard Lacerta lepida 5.20. Levels of endemism of freshwater fish of the large peninsulas of the northern part of the Mediterranean basin 6.1. Potential climatic controls on weathering processes 6.2. Gorges of the River Hérault in south-west France 6.3. Comparison of soil depths as measured in different climate zones and on different lithologies on the island of Lesvos 6.4. Processes leading to the formation of iron oxides and thus the development of brown (goethite) and red (haematite) soils 6.5. Spatial distribution of soils in the Mediterranean basin according to the FAO classification scheme 6.6. Examples of typical soil profiles from the Mediterranean 6.7. Examples of splash pillars forming on marls (the ‘Terres Noires’) near Propiac, south-east France 6.8. Examples of the development of patchy vegetation representing ‘islands of fertility’, from the Montpellier Garrigue, southern France

xix

118

121 140 141 143 144 145 147 148 149

150 151 152 153 153 154 155 156 157 158 159 159 170 172 175 176 177 178 180 181

xx

List of Figures

6.9. Measured soil displacements as a result of tillage erosion using a mouldboard plough in Tuscany 6.10. Accelerated erosion related to slope-parallel tillage in vineyards in southern France, following the storms of 12–13 November 1999 6.11. Soil production by isopods and porcupines on hillslopes at Sede Boquer, southern Israel 6.12. Comparison of relative interrill erosion rates as a function of amount and type of rock-fragment cover of the soil surface 6.13. Aerial photograph of Stipa tenacissima vegetated slopes at the Rambla Honda, southern Spain 6.14. Rills formed during the extreme rainfalls of 22 September 1992 in south-eastern France 6.15. Badlands at Tabernas, Almería, Spain 6.16. Map of observed pipes showing the extent of subsurface erosion in the Agri basin, southern Italy 6.17. Linked pipe and gully erosion, Murcía, Spain 6.18. Map of erosion and deposition following an extreme storm event of 215 mm in Catalonia, Spain in June 2000 6.19. Examples of different types of mass movement 6.20. Landslide inventories for areas in central Italy 6.21. Large rock slides in the valley of the River Guadalfeo, Granada, Spain 6.22. Idealized evolution of ground-surface properties and surface instability 6.23. Analysis of landslide events that resulted in fatalities in Italy 6.24. Details of the Vaiont landslide disaster 7.1. Mediterranean vegetation communities and trajectories of change 7.2. Bioclimatic life zones of the Mediterranean region 7.3. Invasion of sclerophyllous maquis vegetation into an old olive orchard, Crete 7.4. Typical garrigue vegetation in Crete 7.5. Extent of maquis communities across the Mediterranean region 7.6. Distribution maps for circum-Mediterranean taxa, Olea europea subsp. oleaster, Arbutus unedo, Cistus salvifolius, Lavandula stoechas, and vicariant taxa, Quercus ilex and Q. calliprinos, Pinus halepensis, P. brutia, Quercus suber, and Cercis siliquastrum 7.7. Cistus ladanifer-dominated scrub vegetation of the Algarve, Portugal 7.8. Theoretical degradation and regeneration sequences for primary maquis or sclerophyllous evergreen shrub communities 7.9. The typical ‘hedgehog’ shape of the alpine, Euphorbia acanthothamnos, growing among rocky scree of the Psilorites Mountains, Crete 7.10. Olive terraces on Crete 7.11. Basal regrowth of Arbutus unedo in the spring of 2004, following fire in the summer of 2003, Monchique, southern Portugal 7.12. A goat browsing on Quercus coccifera, Psilorites, Crete 7.13. Spread of evergreen sclerophyllous vegetation across the Mediterranean region during the Holocene 8.1. An upland river catchment in the mountains of north-west Greece 8.2. The water balance of the Mediterranean region showing the major fluxes between the main components of the hydrological cycle 8.3. Total annual runoff from river basins in each country bordering the Mediterranean Sea

182

183 184 185 186 187 187 188 189 190 191 192 193 194 195 196 204 205 207 208 209

210 211 212 213 215 219 221 224 230 231 232

List of Figures

8.4. Eagleson’s decomposition of the annual water balance for catchments in different climatic settings 8.5. Rainfall regimes in the Mediterranean region 8.6. The seasonally dry gravel-bed channel of the Voidomatis River upstream of the Vikos Gorge in north-west Greece 8.7. Average monthly flows for rivers around the Mediterranean basin 8.8. The relationship between relative runoff and vegetation canopy cover 8.9. The species used by Garcia-Ortiz in her study of rainfall partitioning by different Mediterranean plants 8.10. Flood flows and erosion in Mediterranean catchments 8.11. Stone-walled terraces on hillslopes near Campanet in central Majorca 8.12. A water cistern directly under the former Greek agora in Ptolomais, Libya 8.13. (a) A newly built check dam in the Sugura River basin, south-east Spain, (b) a sediment-filled reservoir, Valdeinfierno, Murcia, Spain, and (c) oblique air photograph of the town of Puerto Lumbreras and the Rambla Nogalte in the aftermath of the large flood in October 1973 8.14. Map of the Segura River basin (18,800 km2 ), and the impact of reservoir construction on the monthly distribution of flows in the Segura River, Murcia, south-east Spain 8.15. The Río Aguas at Urra in the Sorbas basin, Almeria, south-east Spain, in flood and dry 8.16. Suspended sediment yield from river basins in mountain environments in different climate and vegetation zones 8.17. (a) Gully erosion in soft sediments in central Israel, and (b) a veneer of fresh suspended sediments deposited within the channel zone of the Torcicoda River, central Sicily 8.18. Suspended sediment yield from Moroccan river catchments formed in different rock types 9.1. Maps showing: (a) Mediterranean type climates; (b) the location of the largest lakes in the circum-Mediterranean region; (c) exemplar lake types; (d) location of selected key Holocene palaeolimnological sites 9.2. Lake Pamvotis, Ioannina basin, north-west Greece: an example of a freshwater lake in a karstic intermontane landscape 9.3. Meke Tuzlası , a hypersaline lake occupying a Late Pleistocene crater on the Anatolian plateau, with a new volcanic cone rising through the middle 9.4. Skadar, the largest freshwater lake in the Balkans 9.5. Lago di Pergusa, a small shallow circular lake on Sicily 9.6. Ternary diagram showing the major chemical anion composition of fifty-seven inland lakes in Spain 9.7. The Dead Sea 9.8. Comparative oxygen-isotope curves for three East Mediterranean lake records 9.9. Key selected palaeolimnological indicators and inferred record of Holocene lake-level change in the Laguna de Medina, Cádiz, south-west Spain 9.10. Stratigraphic changes in major pollen types from Birket Ram showing mid–late Holocene cultural and vegetation change in the Golan Heights, along with catchment erosion indicated by magnetic susceptibility.

xxi

233 234 235 236 237 238 239 240 241

242

243 245 246

247 248

256 262

262 264 264 265 266 269

270

271

xxii

List of Figures

9.11. Lakes with contrasting catchment land uses and aquatic ecology in the Middle Atlas region of Morocco: (a) Tigalmamine and (b) Sidi Ali 9.12. Twentieth-century catchment land use and diatom algae species composition from the recent sediments of five Middle Atlas lakes 9.13. (a) Numbers and areas of Mediterranean–Black Sea Ramsar sites by country, and (b) location of selected coastal wetlands and other threatened sites 9.14. The Huleh basin, Israel, showing wetland extent before and after mid-twentieth-century drainage reclamation 9.15. Burdur, a saline lake occupying a tectonic basin in south-west Turkey 9.16. Late twentieth-century urban expansion of Istanbul into the catchments of the Çekmece coastal lagoons 10.1. The distribution of the major outcrops of carbonate rocks (limestones, dolomites, and marble) in the Mediterranean region 10.2. Steep limestone slopes in Kotor Bay on the coast of Montenegro 10.3. Bedding planes and joints in Mediterranean limestones 10.4. Large-scale solution channels in limestone in a formerly glaciated valley, Durmitor Massif, Montenegro 10.5. Geography students from the University of Manchester exploring the cave systems in the gypsum of the Sorbas basin in south-east Spain 10.6. The tectonic setting for the deposition and deformation of Mediterranean limestones 10.7. The development of vadose and phreatic cave systems in a karst drainage system 10.8. A typical karst system in the Mediterranean region showing material inputs, stores, and outputs and associated processes in the vadose and phreatic zones 10.9. Features produced by the precipitation of calcium carbonate in karst environments in Majorca, Spain 10.10. Fine-grained sediment outputs from a Mediterranean karst system 10.11. The vertical zonation of karst landscapes and processes in ‘European folded mountain regions’ 10.12. Two forms of intentional human modification to hillslopes in karst environments in the Dinaric region 10.13. (a) The distribution of poljes in the Dinaric karst region. (b) The formation of three types of poljes under varying structural and hydrogeological conditions 10.14. Two limestone gorges at different stages of development in the Mediterranean region 10.15. Relict karst pinnacles in the White Desert of Egypt 10.16. Three landscape features in the Mediterranean produced by the precipitation of secondary carbonates 10.17. Karst terrain in north-east Majorca showing bare limestone slopes and thick terra rossa soils on the valley floor 10.18. Rockshelter and cave entrance environments can form important sediment sinks and represent a major archaeological resource 10.19. A high resolution oxygen isotope record (‰18 O) from speleothems in Soreq Cave, Israel, spanning the last 140,000 years 10.20. Human activities in the Mediterranean region and their potential impact on non-karst and karst terrains

273 273

275 277 278 279 288 289 291 292

293 294 296

297 299 300 302 303

304 306 307 308 309 311 313 314

List of Figures

11.1. A building under excavation at the archaeological site of Olympia in the valley of the Alfios River in the western Peleponnese, Greece 11.2. (a) The Quaternary sediments of the Louros Valley in Epirus, north-west Greece. (b) The temporal record of channel and floodplain deposition and incision 11.3. Large boulders on the bed of the channel of the Voidomatis River in the Vikos Gorge of north-west Greece 11.4. (a) View looking across the coastal plain in Cyrenaica (north-east Libya) showing the lower course of the Wadi Zewana. (b) Block diagrams showing the evolution of the Late Pleistocene and Holocene alluvial stratigraphy at a trunk stream tributary confluence in wadi systems in Libya 11.5. (a) The lower course of the Wadi Zewana showing a c.25 m thick exposure in Late Pleistocene alluvium. (b) A section showing the Late Pleistocene river sediments with coarse-grained angular gravels exposed at the base 11.6. Dated alluvial units in river systems across the Mediterranean region between c.130 and 10 ka shown in association with two proxy climate records 11.7. Dated alluvial units in river systems across the Mediterranean region between c.65 and 10 ka 11.8. The Pleistocene and Holocene fluvial stratigraphy in the middle and lower reaches of the Voidomatis River basin 11.9. The sediments exposed during the excavations at Boila rockshelter in the lower reaches of the Voidomatis River in north-west Greece 11.10. The deeply incised valley floor and Quaternary terraces of the Río Aguas in south-east Spain 11.11. The Holocene alluvial sediments and terraces in the middle reaches of the Torcicoda River in central Sicily 11.12. The prosperity and depression model of slope stability and soil erosion 11.13. Patterns of Holocene alluviation in Greece for the last 8,000 years 11.14. Patterns of fluvial aggradation and flooding in five Mediterranean countries for the Holocene period 11.15. (a) A summed probability plot for radiocarbon dates from Holocene alluvial records in Spain (11 ka to present). (b) Probability plots for each of the three alluvial depositional contexts shown. (c) Summed probability plots based on the change and mid-point alluvial data sets and the bracketed slackwater flood deposits. (d) A summed probability plot for the radiocarbon dates from the slackwater sediments shown in relation to the North Atlantic drift ice index 11.16. Probability difference curves of radiocarbon dates associated with major flooding episodes in Great Britain, Spain, and Poland 11.17. The flood record for the Aradena Gorge in south-west Crete between 1840 and 2000 based on lichen dating of coarse-grained flood deposits 11.18. Flood histories from five parts of the Mediterranean since AD 1500 11.19. The deeply incised channel zone of the Alfios River in western Greece 11.20. Summary of human impacts on the river channel systems in the lower reaches and delta complex of the Axios, Aliakmon, and Gallikos rivers in north-east Greece over the last century

xxiii

320

322 324

326

327

327 328 332 333 334 336 337 338 339

341 342

343 345 346

347

xxiv

List of Figures

12.1. Map of the Mediterranean showing the main mountain areas referred to in Chapter 12 12.2. (a) Snow accumulation to a depth of c.3 m between the villages of Zabljak and Crna Gora in Montenegro. (b) Looking west from the summit of Mount Orjen towards the summit of Subra, Montenegro 12.3. Map of the Zeleni Sneg glacier on Mount Triglav, Slovenia, depicting ice retreat since the mid-nineteenth century 12.4. The Debeli Namet glacier on the northern slopes of Sljeme, Montenegro 12.5. The effects of avalanching on beech trees on Mount Tymphi, Greece 12.6. The distribution of currently glacierized areas in the Pyrenees 12.7. Unglaciated periglacial surface covered in felsenmeer, Ouanoukrim, Atlas Mountains 12.8. Map of the Glacier d’Ossoue, the largest modern glacier in the Pyrenees 12.9. Map of the glaciers of the Maladeta massif where some of the largest modern glaciers in the Pyrenees are found 12.10. Map of the former extent of Pleistocene glacial and nivation features in the Mediterranean 12.11. Moraines at c.1,700 m a.s.l. in the Vourtapa valley above the village of Skamnelli on Mount Tymphi, Greece 12.12. Limestone pavement on Mount Tymphi, Greece 12.13. Cemented till on Mount Tymphi, Greece 12.14. Well-developed screes within the limits of Vlasian Stage glaciers on the southern slopes of Mount Tymphi, Greece 12.15. The extent of Middle and Late Pleistocene glaciers on Mount Tymphi, Greece 12.16. Glacial geomorphological maps of Mount Olympus, north-eastern Greece 12.17. Moraines at c.1,000 m a.s.l. in Duboki Do, above the village of Ubli on Mount Orjen, Montenegro 12.18. Glacial arête between the peaks of Sedlena Greda and Ranisava in the Durmitor mountain area, Montenegro 12.19. Summary pollen percentage curves from the Ioannina I-284 sequence in north-west Greece, spanning the Last Glacial cycle 13.1. Major tectonic structures and associated seismicity of the Mediterranean region 13.2. Coastal morphodynamics of the Mediterranean basins showing the general near-surface water circulation pattern and the locations and attributes of the four major delta shelves 13.3. The Nile littoral cell extends along the south-eastern Mediterranean coast from Alexandria, Egypt, to the Akhziv Submarine Canyon, Israel 13.4. (a) Elevation of the Last Interglacial shoreline. (b) Rates of late Holocene crustal movement. (c) Predictions of global isostatic adjustment made for Mediterranean tide-gauge stations 13.5. Predicted relative sea levels and shorelines across the Mediterranean region at four epochs 13.6. Schematic representation of a littoral karst cave 13.7. (a) Effects of variations of coastal-wave energy on marine-notch formation. (b) A marine notch at Capo Milazzo in north-eastern Sicily 13.8. The absence of a Holocene sea level above present datum is supported by the evidence of painted horses on a wall of a half-submerged Palaeolithic cave near Marseilles 13.9. A schematic coastal profile showing the main characteristics of bioconstruction and biodestruction on calcareous coasts in the western and eastern Mediterranean

354

356 357 358 359 360 361 363 364 365 366 367 368 370 371 372 373 374 377 387

388 389

391 392 394 395

395

396

List of Figures

13.10. Age-depth diagram from Marseilles’s archaeological excavations compared with dated algal rims from nearby rocky cliffs 13.11. Measured relative sea-level changes in the old harbour of Pozzuoli compared to estimated relative sea-level changes using biological indicators 13.12. Recent relative sea-level variations in Antikythira island, Greece 13.13. Tsunami activity in the Mediterrenean Sea 13.14. Franchthi Cave in the south-east Peloponnese, Greece 13.15. Historical records of coastal flooding for the Rivers Tiber and Rhône 14.1. A map of some aeolian phenomena and locations in the Mediterranean basin 14.2. The passage of dust systems from North Africa to the Middle East, mid-March 1998 14.3. The TOMS sequence across North Africa to the Middle East for mid-April 2000 14.4. The loess of Matmata, southern Tunisia, has been inhabited by cave dwellers 14.5. Barchan dunes in the Libyan Desert, Kharga depression, Egypt 14.6. Gypsum crust soils with polygonal structures in southern Tunisia 14.7. The great lunette dune on the lee side of the Sebkha el Kelbia, central Tunisia 15.1. Map of the Mediterranean basin showing the locations of selected volcanoes and volcanic provinces 15.2. Volcanic hazards: Mt. Etna erupting in 2001 15.3. Fumarolic and diffuse soil emissions on Vulcano (Italy) pose a health hazard 15.4. Map to show the extent of fallout from the Y5/Campanian Ignimbrite eruption 15.5. View of Herculaneum and modern Ercolano 15.6. Plaster cast of one of Pompeii’s victims 15.7. Sequence of laminated deposits from pyroclastic density currents in the Monte Guardia area of Lipari 15.8. Stromboli volcano 15.9. Volcano seen from Lipari 15.10. Aerial view of Mt. Etna rising above the city of Catania 15.11. The town of Fira clinging to the rim of Santorini’s caldera 15.12. Map of part of Santorini in c.1715, showing the Kameni islands after the 1707 eruption 15.13. (a) Map of Nea Kameni and Mikra Kameni, after the 1866–70 eruptions, and (b) hillshade digital elevation model of the present-day status of the Kameni islands 15.14. (a) The duration of eruptions on the Kameni islands since AD 1570, and (b) the heights of lava domes as a function of time elapsed during an eruption for the 1866 and 1939 Kameni eruptions compared to the domes of Mt. St Helens and St Vincent 15.15. Volcano monitoring and crisis response 16.1. (a) Ruins of buildings in Verneuges, Provence, that were destroyed by the 1909 Lambesc earthquake. (b) A tombstone from the ancient Greek town of Nikomedia, commemorating the death of two young boys and their teacher in AD 120

xxv

398

399 399 400 403 404 416 419 420 421 422 423 424 436 438 440 443 445 446 447 448 448 449 451 452

453

455 458

470

xxvi

List of Figures

16.2. Epicentres of shallow earthquakes in the Mediterranean, 1961–83 16.3. Epicentres of earthquakes, 1900–65, and areas affected by historical earthquakes in the eastern Mediterranean 16.4. Plate boundaries in the eastern Mediterranean 16.5. Seismicity across the Peloponnese 16.6. Examples of earthquake damage to ancient buildings 16.7. Faulting and river response during the 1980 earthquake in southern Italy 16.8. Rocks uplifted during the 1953 earthquakes, Ionian Islands 16.9. Progressive rupturing of the North Anatolian Fault, 1939–99 16.10. Areas affected by the 1202, 1926, and 1927 earthquakes, eastern Mediterranean 16.11. (a) A normal fault produced by the 1954 earthquake in Thessaly. (b) Reverse faulting, folding, and uplift that dammed the Cheliff River during the 1980 Al Asnam earthquake, Algeria. (c) Railway tracks offset by strike-slip faulting during the 1999 Izmit earthquake in Turkey 16.12. Contours of uplift across western Crete resulting from the AD 365 event 16.13. Response of a stream at Sougia, Crete, to an uplift during the AD 365 earthquake 16.14. The effects of ground sliding, compaction, and perhaps liquefaction following the 1783 earthquake in Calabria, Italy 16.15. (a) A collapsed multi-storey building in Kalamata, Greece, following the 1986 earthquake. (b) During the same event many traditional buildings were badly damaged but did not collapse 17.1. Co-seismic dip-slip motion along faults and tsunamigenesis near deep sea trenches 17.2. A schematic explanation of some of the tsunami terms used in Chapter 17 17.3. (a) Tsunamigenic zones in the Mediterranean Sea. (b) Types of field evidence for the occurrence of past tsunamis 17.4. Important tsunamis reported for southern Italy 17.5. Some of the damage caused by the Stromboli tsunami of 30 December 2002 17.6. Important tsunamis along the Hellenic arc 17.7. An excavated section showing palaeotsunami deposits in Dalaman, south-west Turkey 17.8. Important tsunamis in the Cyclades Islands in the southern Aegean Sea 17.9. Palaeotsunami investigation within an archaeological excavation in St George, in eastern Thera 17.10. Detail of the tsunami deposits exposed and attributed to the 30 September 1650 volcanigenic tsunami 17.11. (a) Important tsunamis in the Gulf of Corinth and the Maliakos Gulf. (b) The coast prior to the large landslide that produced a tsunami in the Gulf of Corinth, 1963 17.12. Important tsunamis in (a) the east and north Aegean Sea, (b) the Sea of Marmara, and (c) the Cyprus-Levantine Sea area 17.13. The frequency of tsunamis in the Mediterranean Sea as a function of longitude and latitude, 1628 BC to AD 2003 17.14. The cumulative frequency and intensity of Mediterranean tsunamis as a function of time, 1628 BC to AD 2003 17.15. Relationship between between intensity and frequency of tsunamis for Greece, Italy, and the Mediterranean Sea

472 473 474 475 476 477 477 478 479

481 482 483 484

486 494 494 496 497 498 499 500 500 501 502

503 504 506 507 507

List of Figures

17.16. Tsunami intensity as a function of earthquake intensity and magnitude for the entire Mediterranean Sea 18.1. (a) The key study areas of adverse weather phenomena in this chapter. (b) The key sites and river systems in north-east Spain and southern France cited in the text 18.2. The mean annual frequency of cyclones in the Mediterranean, summer and winter 18.3. Classification of cloud systems associated with heavy rainfall events in the Mediterranean area 18.4. An example of the synoptic situation that produces heavy rainfall with catastrophic floods in the western Mediterranean region 18.5. Rainfall distribution for a flood event of Type 2b, 6–8 November 1982 18.6. Heavy rainfall that produced a flash flood event of Type 1, 3 April 1989 in the Pyrenees 18.7. A flood event of Type 2a, 10 June 2000, in Catalonia 18.8. Destruction of a road bridge by the Riera de la Magarola during the flood event of 10 June 2000 18.9. Some consequences of the flash flooding produced in the Gard region of southern France, 8 September 2002 18.10. A flood event of Type 3, 21–30 January 1996, in Catalonia 18.11. The relationship between maximum flow and average annual catchment flow for river catchments in the Mediterranean and non-Mediterranean regions of Europe 18.12. The record of catastrophic floods in Catalonia since the early fourteenth century 18.13. A ‘Levante’ wind storm that affected Catalonia, 16–18 October 2002 18.14. The destruction of forests in the Balearic Islands by the western Mediterranean ‘superstorm’ of November 2001 18.15. A tornado recorded in Barcelona, on 8 September 2005 19.1. Fires in Mediterranean basin countries 19.2. Temporal variation in (a) the number of fires per year and (b) the area burnt 19.3. The proportion of fires and of area burnt in relation to fire size, France 19.4. The area burnt by forest fires in Portugal during 2003 19.5. A fire scar at the base of a pine tree trunk in Catalonia, north-east Spain 19.6. A conceptual model of factors influencing the area burnt in a region 19.7. (a) Changes in summer climatic fire risk in Catalonia, and (b) the burnt area in relation to the number of days with high fire risk 19.8. Fire risk in Europe according to the Canadian FWI 19.9. Burned forest in a highly populated area of Catalonia 19.10. A fire prevention sign in the uplands of Majorca 19.11. An illustration of a fire regime model 20.1. Different phases of the Holling and Gunderson adaptive cycle 20.2. An application of the USLE for the Autonomous Region of Andalucia 20.3. The EU erosion estimate for Spain 20.4. Soil loss and runoff as a function of the proportion of ground covered by a vegetation canopy 20.5. A metaphor for a system’s stable and unstable conditions 20.6. (a) The relationship between sediment yield and annual effective rainfall. (b) Estimated rates of erosion by wind and water. (c) Erosion for field plots around the Mediterranean

xxvii

509

516 517 518 520 522 524 526 527 527 528

529 531 532 533 533 542 543 544 545 547 549 549 551 552 554 555 565 567 568 569 569

570

xxviii

List of Figures

20.7. The evolution of plant cover over time on abandoned plots in the Spanish Pyrenees 20.8. Runoff and sediment yield data for different land uses on regenerated abandoned fields in the Spanish Pyrenees 20.9. Simulated response of vegetation to stochastic rainfall 20.10. A nested set of adaptive cycles in time and space 21.1. Distribution of potential annual runoff (effective precipitation) in the Mediterranean basin 21.2. The Mediterranean drainage network and river basins 21.3. Natural renewable and exploitable water resources per country in the Mediterranean basin 21.4. Natural renewable water resources and real exploitable water resources per inhabitant in the Mediterranean basin 21.5. The exploitation index of natural renewable water resources across the Mediterranean basin 22.1. European air pollution emissions, 2000 22.2. Widespread aerosol haze in the Mediterranean basin 22.3. Schematic of air flows during MINOS, 2001 22.4. Transport-time spectrum showing the period between release of air pollutants and arrival in the upper troposphere 22.5. Model-calculated ozone at the surface for the present and the possible future 22.6. Mean diurnal cycles of OH and NO3 radicals 22.7. Mean particle composition during MINOS for fine and coarse mode aerosols 22.8. Diurnal mean, clear-sky radiative forcing during MINOS 22.9. (a) Estimated historical SO2 emissions in Europe; (b) 5-year running mean of Mediterranean SST anomalies 22.10. Percentage changes in annual precipitation comparing low and high SST 23.1. The ten regional hotspots of the Mediterranean basin for plant endemism and richness 23.2. A montado in Portugal with cattle and charcoal burners 23.3. Forest recovery on ancient terraces 23.4. The last individual of a former forest of Juniperus thurifera in southern Morocco 23.5. Wetlands, among the most threatened habitats in the Mediterranean region 23.6. The crustacean branchiopod Triops cancriformis with the rare thrumwort Damasonium stellatum and the parsley frog Pelodytes punctatus, Camargue 23.7. A rare amphibian of temporary ponds of the western part of the basin, the Mediterranean newt, Triturus marmoratus 23.8. Urbanization of coastal areas threatens habitats and rare plants and animals 23.9. (a) Mean laying date of the Blue tit Cyanistes caeruleus in mainland and Corsican habitats (b) Variation of clutch size in the Corsican and mainland habitats 23.10. Intense habitat degradation due to repeated fire events results in very low scrubby vegetation and bare ground 23.11. Carpobrotus acinaciformis, a very aggressive invasive plant species in coastal areas

573 573 574 578 584 585 588 589 593 600 601 602 604 605 607 608 609 609 610 616 617 619 620 622

623 624 625

626 627 630

List of Figures

23.12. Ramet of Zelkova sicula (Ulmaceae), a relict and very threatened palaeoendemic small tree, south-eastern Sicily 23.13. Acis fabrei, a narrow endemic with only four known populations from the southern slopes of the Mont Ventoux 23.14. Three taxa of salmonids: Marble trout, Corsica trout, and Brown trout 23.15. The Pond terrapin Emys orbicularis 23.16. A pair of Bonellis’eagles Hierraaetus fasciatus at their nest 23.17. The Scops owl Otus scops, threatened by the decline of large invertebrates 23.18. The Vikos Gorge with its spectacular limestone cliffs 23.19. Cliffs are important habitats for several rare birds and endemic plant species throughout the Mediterranean basin 23.20. Distribution maps within the Mediterranean bioclimatic region of fifty glacial refugia

xxix

632 633 635 636 637 639 641 644 645

List of Tables 1.1. A summary of the impact of tectonics on the geomorphology of the Mediterranean region 1.2. Examples of badlands and partial badlands found across the Mediterranean 2.1. Contributions to total Nile discharge from the main tributaries 3.1. Centres of cyclogenesis in the Mediterranean 3.2. Local winds of the Mediterranean 3.3. Decadal distribution of heatwave days 3.4. Summary of some recent flash floods in Mediterranean Europe 4.1. Scales of environmental variability and vegetation responses 4.2. List of genera of species’ first fossil appearance in the Tertiary record 4.3. A sample of Mediterranean pollen records 5.1. Endemism rates of several groups of species in the Mediterranean 6.1. Comparison of porportions of different soil types according to the FAO classification relative to location in the north or south of the Mediterranean basin 7.1. Terms used to describe Mediterranean sclerophyllous shrubland 7.2. Fire-adapted strategies of some selected Mediterranean taxa 7.3. Flammability of selected Mediterranean plants based on laboratory tests of leaf ignition of Cretan species 8.1. Water and sediment fluxes from the Axios and Aliakmon rivers that drain into the north-west Aegean Sea 9.1. Individual characteristics of all permanent natural lakes in the circum-Mediterranean region >200 km2 in area, excluding coastal lagoons 9.2. Exemplars of Mediterranean lake types 9.3. Selected key Holocene palaeolimnological records for the Mediterranean 9.4. Selected key Mediterranean wetlands requiring conservation or restoration 10.1. Large discharge springs of the world with flows >20 m3 s−1 10.2. The twenty deepest caves in the world 10.3. Characteristics of active karst settings and passive karst settings for rockshelter and cave entrance environments in limestone terrains 10.4. Karst sites in the Mediterranean region with World Heritage status 12.1. Modern glaciers in the Mediterranean 12.2. Correlation table showing the relationship between the fragmentary glacial sequence in the Pindus Mountains, Greece, and the continuous lacustrine parasequence in the nearby Ioannina 249 and 284 cores 12.3. Current understanding of the geochronology of glacial deposits in the Mediterranean region 13.1. Coastal environments around the Mediterranean Sea classified into bedrock and accretion coasts 13.2. Amplitude, duration, permanence, and length of coast affected by various types of rapid relative sea-level change in the Aegean 14.1. Dust over the Mediterranean 14.2. Dust deposition amounts across the Mediterranean

7 19 40 75 78 79 86 90 98 109 158

177 206 218 220 246

257 258 259 260 298 301 310 314 354

369 376 386 401 407 421

List of Tables

14.3. Particle size characteristics of dust in various parts of the Mediterranean 15.1. Major ash layers correlated with known volcanic eruptions in the Mediterranean over the past 200 kyr 15.2. Summary of volcanic hazards 15.3. Twentieth-century record of deaths, injuries, and other impacts of volcanic activity in the Mediterranean 15.4. Fatal eruptions of Somma-Vesuvio 16.1. Some major or otherwise noteworthy earthquakes in the Mediterranean 16.2. A list of indicative criteria for the identification of earthquakes from archaeological data 17.1. Strong tsunamis of intensity k ≥ 4 reported for the Mediterranean Sea between 426 BC and AD 2002 17.2. Mean return period of tsunami intensity and the most likely maximum tsunami intensity to be observed for various parts of the Mediterranean Sea 17.3. Tsunami potential in each of the tsunamigenic zones of the Mediterranean 18.1. Major flood events in the European Mediterranean since 1990 18.2. The number of catastrophic floods based on historical sources recorded in various river basins in Spain, Italy, and France 20.1. National soil erosion risk data for five Mediterranean countries in the EU 21.1. Water resources in the Mediterranean basin by country and continent 21.2. Key figures on internal and external natural and exploitable water resources for the three main regions of the Mediterranean basin 21.3. Annual water withdrawal volumes for the three main regions of the Mediterranean basin 21.4. Present-day pressures on water resources in the Mediterranean basin 21.5. Water demand predictions for 2025 in the three regions of the Mediterranean basin 22.1. Air pollution emissions in Europe in the year 2000, and emission reductions between 1980 and 2000 23.1. Some of the most invasive alien plants occurring in the Mediterranean basin 23.2. Threatened vascular plants by country based on the former IUCN categories and included in the 1997 IUCN Red List of Threatened Plants 23.3. Threatened vascular plants on the seven large Mediterranean islands, based on the former IUCN categories 23.4. List of large mammals that were present in the Mediterranean basin during the Late Pleistocene, including species found as fossils in various deposits of southern France, and that became extinct 23.5. Major protected areas such as National Parks and Biosphere Reserves within the Mediterranean bioclimatic region 23.6. List of Ramsar sites within the Mediterranean bioclimatic region 23.7. Impacts of the major influences on the biodiversity of ten Mediterranean regional hotspots

xxxi

421 436 437 441 444 470 475 495

507 508 514 530 566 587 588 593 594 595 600 629 631 631

634 640 640 644

List of Contributors Ramadan Abu-Zied is an assistant lecturer in the Geology Department at Mansoura University in Egypt. He obtained his Ph.D. in 2001 with Eelco Rohling in Southampton. He is a specialist in benthic foraminifera and the application of their abundance variations and stable isotope ratios in palaeoenvironmental research, with emphasis on the eastern Mediterranean. Harriet Allen is a lecturer and researcher in the Department of Geography, University of Cambridge. Her research interests focus on the response of ecosystems to environmental change. This includes the integration of high resolution remote sensing data with ecological surveys to assess contemporary ecosystem changes, and pollen and sedimentological research to reconstruct longer-term ecosystem changes. Much of her recent fieldwork has been carried out in Mediterranean-climate regions, including Greece and Portugal. Jacques Blondel is with the Centre d’Écologie Fonctionnelle et Evolutive, CNRS, Montpellier and he previously taught at the University of Louvain in Belgium. His main research interests focus on the origin and regulation of biological diversity at several scales of space and time, from processes involved in the establishment of faunas at the scale of the Mediterranean region, to community dynamics at the scale of landscapes and populations at the scale of local habitats. He conducts a long-term (>30 years) programme on the phenotypic variation of birds in Mediterranean habitat mosaics. He is also concerned with biodiversity and conservation issues. Maria del Carmen Llasat is a professor and coordinates a research group in the Department of Astronomy and Meteorology at the University of Barcelona. She was president of the Natural Hazards Section in the former European Geophysical Society (now part of the European Geosciences Union). She is the managing editor of the journal Natural Hazards and Earth System Science. Her research interests are mainly meteorological and hydrometeorological risks in the Mediterranean region. James Casford in a lecturer in the Department of Geography at Durham University. He obtained his Ph.D. in 2001 with Eelco Rohling in Southampton. His research focuses on climate variability and the palaeoceanography of marginal basins, particularly the eastern Mediterranean. Marc Castellnou is the Fire Analysis Officer in the Fire Service of the Catalonian Regional Government where he runs the forest fire training programme for forest fire professionals. He has worked and tackled fires in the USA, Africa, France, the UK, and Portugal. He has also carried out research into forest fire regimes, forest fire ecology, and forest fire propagation. He has been involved in several European research projects as researcher and coordinator. More recently he has coordinated the operational training of international fire-fighting units. Andrew Goudie is a Professor of Geography in the University of Oxford and Master of St Cross College. His research interests are in desert geomorphology and climate change. He has worked extensively in Africa, India, and the Middle East on such themes as dunes, dust, pans, loess, and salt. Andrew Harding is a research associate at the Global Environmental and Climate Change Centre (GEC3 ) at McGill University working with Environment Canada. His main focus is the investigation of linear and non-linear links between synoptic scale atmospheric dynamics and meso-scale climate extremes. His Ph.D. (from the Climatic Research Unit

List of Contributors

xxxiii

at the University of East Anglia) focused on patterns and trends evident within the variability and sensitivity of Mediterranean climate extremes. Angela Hayes is with the Department of Geography, Mary Immaculate College at the University of Limerick. She obtained her Ph.D. in 1999 with Eelco Rohling in Southampton. She specializes in reconstructing past sea surface temperatures from planktonic foraminiferal abundance data, and uses this in combination with stable isotope ratios to reconstruct past ocean and climate conditions. Tom Holt is a Senior Research Associate in the Climatic Research Unit, University of East Anglia. His main research interests are the assessment of uncertainty in projections of climate change from numerical models and the analysis of climate extremes. A special focus has been on likely changes in climate extremes over the Mediterranean to 2100, with a particular emphasis on drought and impacts on the tourism, agriculture, and energy industries. Babette Hoogakker is a postdoctoral researcher in the Department of Earth Sciences at the University of Cambridge. She obtained her Ph.D. in 2003 with Eelco Rohling in Southampton. She is specialized in the combined use of sedimentology, foraminiferal census counts, and shell chemistry, to reconstruct palaeoceanographic conditions on global scales. Philip Hughes is a lecturer in Physical Geography at the University of Manchester, where he held a postdoctoral fellowship between 2004 and 2006 working on the glacial and periglacial history of the Mediterranean. His Ph.D. (University of Cambridge, 2004) focused on Quaternary glaciation in the Pindus Mountains of north-west Greece and he has also worked in Montenegro and Morocco. Jos Lelieveld is director of the Max Planck Institute for Chemistry, and is professor in Atmospheric Physics at Mainz University. His research addresses photo-oxidants (e.g. ozone), the cleaning mechanism of the atmosphere, aerosols, and links with climate. John Lewin is Emeritus Professor of Physical Geography at Aberystwyth University. He is a fluvial geomorphologist who has been concerned with the development of river landforms over a wide range of timescales: long-term landscape evolution dating back to the Tertiary, Quaternary alluvial deposits, historical river channel changes, and contemporary river processes. Research over the more recent timescales has especially involved human impacts and the dispersal of polluted sediments. He has wide field experience of the Mediterranean region. He co-edited Mediterranean Quaternary River Environments (1995) with Mark Macklin and Jamie Woodward. Francisco Lloret is Professor of Ecology at the Universitat Autònoma Barcelona, and researcher at the CREAF (Centre for Ecological Research and Forestry Applications, Spain). His research interests focus on the structure and dynamics of plant communities in relation to anthropogenic sources of disturbance, such as fire regime, land use change, climate change, and exotic plant invasion. He has studied the historical patterns of fire regime in the Mediterranean basin and the interaction between fire regime and vegetation recovery after fire. He has worked in the Mediterranean basin, the United States, Mexico, and Australia. Francisco López-Bermúdez is a Professor in the Department of Physical Geography at the University of Murcia. He published extensively on the hydrology and dynamics of fluvial systems in semi-arid environments. He is especially interested in the interactions between vegetation and erosion in Mediterranean river basins and in the generation and geomorphological impact of large floods. In 2006 (with Jorge García-Gómez) he published Desertification in the Arid and Semiarid Mediterranean Regions. A Food Security Issue in the NATO Security through Science Series.

xxxiv

List of Contributors

Mark Macklin is Professor of Physical Geography at Aberystwyth University. His principal research interest is studying fluvial system responses to short- and longer-term environmental change. His work on Mediterranean rivers has focused on the Iberian Peninsula, mainland and island Greece, and Libya. These investigations have documented the effects of Late Pleistocene glaciation and sub-orbital scale climate change on river behaviour, established the geomorphic impacts of extreme Holocene flood events on mountain catchments and bedrock gorges, and developed a range of remediation and management strategies for river basins contaminated by historical and present-day metal mining. He has just started a three-year project investigating the role of tectonics on historical river development in western Crete. Jean Margat is a geology graduate who worked as a hydrogeologist from 1947 to 1989, first in Morocco (Geological Survey), then in the ‘Bureau de Recherches Geologiques et Minières’ (BRGM: French Geological Survey) focusing on groundwater research, in France and in the arid zone. Later he specialized in water resources in the Mediterranean basin. He is the author of numerous publications dealing with the assessment and the management of water resources, the most recent being ‘Water for the Mediterranean, Present and Future’ (Mediterranean Action Plan /Blue Plan, 2004). He is vice-chairman of the Association ‘Blue Plan for the Mediterranean’ and the ‘Mediterranean Institute of Water’. Anne Mather is a Reader in Earth Sciences with the School of Geography at the University of Plymouth. Her research focuses on long-term landscape development in drylands with the main focus on tectonic geomorphology. This research encompasses both direct and indirect responses of alluvial and fluvial systems to regional tectonics. Her main geographical areas of research include Spain, Turkey, Morocco, and Chile. Frédéric Médail is Professor of Plant Ecology and Biogeography in the Mediterranean Institute of Ecology and Palaeoecology (IMEP, University Paul Cézanne Aix-Marseille III). His research interests include the conservation and biogeography of Mediterranean plants, the processes induced by biological invasion and insular ecology. He conducts his research at several ecological scales, from regional phylogeography to the population biology and ecology of rare and endemic plants better to understand patterns and processes involved in the diversity of the Mediterranean basin hotspot. Christophe Morhange is Professor of Physical Geography at the University of Provence and a member of the CEREGE’s (CNRS) Geomorphology and Tectonics group, Aix-enProvence, France. His research interests are: (1) Holocene relative sea-level changes using biological indicators; and (2) coastal geoarchaeology, notably the use of ancient harbour archives to reconstruct natural and anthropogenically forced changes since antiquity. He has worked in numerous sites around the Mediterranean and the Black Sea, including Bulgaria, Cyprus, Egypt, France, Greece, Israel, Italy, Lebanon, Spain, Tunisia, and the Ukraine. Clive Oppenheimer is based at the Department of Geography, University of Cambridge. His research focus is the development and application of remote sensing techniques for environmental monitoring, especially volcanology. His observations of volcanic gas and aerosol emissions have been used to investigate the transport of magma below volcanoes, as well as the impacts of volcanic pollution on the atmospheric environment. Recently, he has studied the lava lakes of Mt. Erebus in Antarctica, and Erta ’Ale in Ethiopia. He is also interested in the climatic and human impacts of major historic and prehistoric eruptions. Jean Palutikof is Head of the Technical Support Unit, IPCC Working Group II (Impacts, Adaptation, and Vulnerability). She is based in the Hadley Centre at the UK Met Office. Previously, she worked in the Climatic Research Unit at the University of East Anglia, and in the Department of Geography at the University of Nairobi. Her research interests

List of Contributors

xxxv

focus on climate change impacts, and the application of climatic data to economic and planning issues. She worked on the EU-funded MEDALUS projects, constructing scenarios of regional climate change for the Mediterranean region. Gerassimos Papadopoulos is Research Director with the Institute of Geodynamics, National Observatory of Athens, Greece. His main research interests are in instrumental and historical seismicity, earthquake prediction, and tsunamis, particularly in the EuroMediterranean region. He has worked as a visiting scientist at MIT (Boston, 1984), NIED (Tsukuba, Japan, 1993), and Tohoku University (Japan, 2004. He served as President of the International Natural Hazards Society (2000–6) and he is Vice-President of the European Seismological Commission. Josep Piñol is at the Centre for Ecological Research and Forestry Applications (CREAF) at the Autonomous University of Barcelona. His first work on forest fires focused on the development of methodologies to estimate fire risk, in particular using meteorological indices and measurements of the moisture content of fine fuels. More recently, he has focused his attention on understanding fire regimes in Mediterranean regions, and, in particular, the role of fuel build-up in increasing the occurrence of very large fires. David Pyle is Professor of Earth Sciences in the Department of Earth Sciences at Oxford University and previously taught at the University of Cambridge (1991–2006). His principal interests are understanding patterns and processes of active volcanism, in particular the dispersal of tephra during large eruptions; the emission and reactivity of gases from volcanic vents; and the interactions between volcanoes and the climate system. He has worked in Greece, Italy, Russia, the Americas, and south-east Asia. Jane Reed is a lecturer in the Department of Geography, University of Hull. Her research interests focus on the use of lacustrine diatoms as quantitative environmental indicators in fresh and saline lakes, both in the palaeolimnological study of lake sediment cores and in environmental biomonitoring. Her main geographical focus is the lakes of the Mediterranean and the Balkans, encompassing research themes ranging from long-term Quaternary climate change to recent water pollution and desiccation. Neil Roberts is Professor of Physical Geography at the University of Plymouth. He received his Ph.D. from the University of London (UCL), and has been a researcher at the University of Oxford and subsequently Lecturer at Loughborough University. His research emphasizes global change during the Late Quaternary period, specifically lake sediment-based archives of past climate variability in low and mid-latitude regions, with links to archaeology. He is author of the key text, The Holocene, published by Blackwell. Eelco Rohling is a Professor at the School of Ocean and Earth Science, Southampton University, and is based at the National Oceanography Centre, Southampton. He works on the processes of ocean and climate change over a range of timescales, with emphasis on Pleistocene and Holocene records from subtropical marginal seas, such as the Mediterranean and the Red Sea. Iain Stewart is Professor of Geoscience Communication at the University of Plymouth. A former president of the INQUA Commission on Neotectonics, his primary research interests are on Holocene coastal tectonics and sea-level change, with particular emphasis on shoreline records of earthquake, volcano, and tsunami activity. With Claudio VitaFinzi, he co-edited The Geological Society of London’s Special Publication 146 on Coastal Tectonics, and has studied Holocene coastal change in tectonically active parts of the Mediterranean, principally Aegean Greece and Turkey, and eastern Sicily. Stathis Stiros is Associate Professor with the Department of Civil Engineering, Patras University, and was previously with the Institute of Geology and Mineral Exploration (IGME)

xxxvi

List of Contributors

in Athens. His research interests include the identification of earthquakes from archaeological and coastal geomorphological data and the modelling of crustal and ground deformation processes. He is also interested in new developments in geodetic instrumentation and analytical tools for the recording and analysis of small-scale oscillatory movements, as well as geodetic techniques used in antiquity. John Thornes, who died in 2008, was Research Chair in Physical Geography at King’s College, University of London. His main research was in the interaction of grazing, vegetation and erosion. He has jointly published Environmental Issues in the Mediterranean with John Wainwright for Routledge. He was awarded Doctor Honoris Causa by the University of Murcia in 2006 for his contributions to the knowledge of the environments of southeast Spain. Chronis Tzedakis is Professor of Global Change Palaeoecology at the Earth and Biosphere Institute, School of Geography, University of Leeds. His research centres on understanding the response of vegetation to variations in climatic forcing on different timescales (orbital and millennial/centennial) in the Mediterranean region. His work involves the study of long lake sequences and deep-sea cores, which provide an opportunity to examine phase and amplitude relationships between climate and vegetation changes over several glacial–interglacial cycles. John Wainwright is Professor of Physical Geography at the University of Sheffield, and was previously at King’s College London. His research focuses on the role of erosion processes in land degradation over a range of timescales from prehistory to the modern day, using combined field, laboratory, and computer-modelling approaches. He has worked extensively in southern France, Spain, Italy, and Greece, as well as in dryland environments in the USA and Africa. Jamie Woodward is Professor of Physical Geography at the University of Manchester. His research focuses on Quaternary environmental change, geomorphological systems, and geoarchaeology in the Mediterranean. He has field experience in various parts of Greece, Sicily, Montenegro, Corsica, and Spain. He has also worked in the Nile Valley with the British Museum. He is the co-editor of Geoarchaeology: An International Journal. He is especially interested in fluvial and glacial archives of change and has also worked on Late Pleistocene and Holocene sedimentary records in Mediterranean rockshelters and caves.

I

The Physical and Biological Framework

This page intentionally left blank

Editorial Introduction Jamie Woodward

By examining both contemporary processes and longterm records of change, this volume explores the climates, landscapes, ecosystems, and hazards that comprise the Mediterranean world. This is the only region on Earth where three continents meet and their interaction has produced a very distinctive physical geography. This book examines the landscapes and processes at the margins of the three continents and the distinctive marine environment between them. In broad terms, the physical geography of the Mediterranean is a product of long-term interplay between tectonic forces, climate change, river basin and marine processes, and biosphere dynamics, as well as the action of humans during the course of the Holocene. From the outset, it is important to keep in mind that this physical geography is an integration of energy, materials, and processes within a much wider global system. The Mediterranean is a zone of convergence and interaction. It is a meeting place not only for tectonic plates, but also for air masses, energy, and river flows from both temperate and tropical latitudes. The region also interacts directly with the global ocean, receiving cool North Atlantic waters in exchange for the warmer and saltier waters produced in the basins of the Mediterranean Sea. It is also a biodiversity hotspot; the Mediterranean has been a meeting place for plants, animals, and humans from three continents throughout much of its history. The chapters in Part I set out the physical and biological framework for the rest of the book and examine key debates about the evolution of the Mediterranean environment. They explore fundamental interactions between the lithosphere, atmosphere, hydrosphere, and biosphere across a range of spatial and temporal scales. The scene is set for later chapters that focus more closely on particular aspects of the Mediterranean environment such as ecosystem dynamics, river basin systems, karst environments, natural hazards, and land degradation.

Chapter 1 examines the role of tectonic processes in the development of the Mediterranean landscape and its marine basins. Also highlighted are the dramatic environmental changes and the geomorphological legacy associated with the Messinian Salinity Crisis of the Late Miocene. Chapter 2 focuses on the marine environment, both ancient and modern. The Mediterranean Sea is a relatively small body of water in the global ocean system that has reacted in a sensitive way to environmental change; this is well illustrated by the repeated formation of sapropels throughout the Quaternary. The climate system is the focus of Chapter 3 and this completes the trio of opening chapters that explore the land, sea, and atmosphere of the Mediterranean and some of the key interactions between them. Chapters 4 and 5 examine major themes and ideas in Mediterranean biogeography by exploring the nature and evolution of the region’s vegetation and vertebrate fauna respectively. Chapter 4 combines an analysis of the long-term record of Cenozoic climate and vegetation change with a detailed evaluation of the region’s long Quaternary vegetation records. Together, Chapters 2 and 4 show how key components of the Mediterranean environment responded to the rhythms of global change during the Quaternary. The region provided important refugial areas for plants, animals, and humans during the cold stages of the Quaternary. Pollen records show that trees expanded and contracted their ranges very rapidly in response to abrupt changes in the climate system. Such ecological changes were part of a highly dynamic Quaternary geography that tracked the major environmental fluctuations of the Northern Hemisphere. These themes are developed further for specific environments and settings in the chapters of Part II in conjunction with assessments of the nature and significance of human activity across the Mediterranean in the Holocene.

This page intentionally left blank

1

Tectonic Setting and Landscape Development Anne Mather

Introduction The Mediterranean is the westernmost part of the globalscale Alpine-Himalayan orogenic belt which stretches from Spain to New Zealand (Figure 1.1a). The landscapes of the region have a long and complex history that includes both horizontal and vertical crustal movements and the creation and destruction of oceans (Table 1.1). This began with the break up of the supercontinent Pangea around 250 Ma, which generated the Tethys Ocean—the forerunner to the present-day Mediterranean Sea. Collision of the African and European tectonic plates over the last 30 Ma led to the destruction of the Tethys Ocean, although a few remnants of its geology are preserved within the eastern Mediterranean. It is the collision of Africa and Eurasia, and the associated tectonics that have been largely responsible for generating the Mediterranean Sea, its subsequent history, and the landscapes that surround it. This collisional history progressively reduced the connectivity of the Mediterranean Sea with surrounding marine bodies by closing and restricting marine gateways. During the Miocene, for example, the Mediterranean basin became completely isolated from surrounding marine bodies in what is known as the ‘Messinian Salinity Crisis’. This period saw major changes to the regional water balance leading to evaporation and draw-down of the Mediterranean Sea. This had profound impacts on all aspects of the physical geography of the region including the climatology, biogeography, and geomorphology and its legacy can be seen across the region today. The more recent Quaternary geodynamics of the Mediterranean have generated an area which includes

a complex mixture of zones of plate subduction of various ages and stages (Figure 1.1b). The modern Mediterranean includes zones of active subduction associated with volcanic activity—such as the Calabrian arc—and older zones of now quiescent subduction such as the Betic-Rif arc. There is a wide range of seismic activity associated with these regions from deep (600 km) to shallow (<50 km) and ranging in magnitude up to 8.0 Mw (earthquake moment magnitude; a quantitative and physically based scale for measuring earthquakes). Together, the steep relief and seismic activity generated by the geodynamics drive many of the active erosional landscape processes within the region, from gully and slope erosion leading to the development of badlands, to large (500 km3 ) submarine mass movement events. The geomorphic expression of the regional tectonics in the landscape thus contains valuable data on the spatial and temporal variations in rates and styles of tectonic processes around the Mediterranean. The geomorphic features include river and marine terrace records and alluvial fan sediments. These topics will be touched upon only in their geodynamic context here, but are discussed in more detail later in this book (e.g. Chapters 11 and 13). This chapter focuses on the geologically more recent evolution of the Mediterranean over the last 10 million years. The geodynamics of the region provide a template for understanding the generation of the larger landscape features such as the mountain ranges and sedimentary basins, and the distribution of key landscape elements across the region. The chapter will begin by examining the geodynamics of the region and then consider how these have impacted upon the long-term development of landscapes. Thus, this chapter underpins many

6

Anne Mather

Fig. 1.1. (a) The Alpine Himalayan orogen (in dark grey) in its global setting (modified from Lister et al. 2001) and (b) the main tectonic landform features of the Mediterranean referred to in the text (modified from Carminanti and Doglionini 2004).

Tectonic Setting and Landscape Development

7

TABLE 1.1. A summary of the impact of tectonics on the geomorphology of the Mediterranean region as discussed in this chapter Approximate timing

Tectonic events

5 Ma to Present

The Calabrian arc continues to migrate eastwards. The Hellenic arc continues to migrate southwards.

6–5 Ma

Closure of Atlantic connection (Betic and the Rifian Corridors) took place just after 6 Ma followed by opening of the Straits of Gibraltar and reconnection with the Atlantic Ocean prior to 5 Ma.

15–10 Ma

The Calabrian subduction zone migrates eastwards by rollback, becoming more arcuate. Dinarides subduction slows down, but Hellenic subduction is very active leading to southerly rollback of the main Hellenic arc. The Carpathians migrate eastwards due to rollback subduction. Calabrian subduction zone develops. The Tethys Ocean is being destroyed through subduction. The Alps and Betics form a continuous subduction zone in the west. The Dinarides, Hellenides, and Taurides belts are associated with active subduction in the east. Breakup of the supercontinent Pangea and creation of the Tethys Ocean.

30 Ma 45 Ma

250 Ma

of the issues related to the Mediterranean environment discussed in this book. These include understanding the geomorphology and dominant process regimes as well as the long-term record of human interaction with the environment (Part II), many of the natural hazards in the region (Part III), and some of the key contemporary environmental issues associated with human use of the Mediterranean environment and its resources (Part IV).

Geodynamics of the Mediterranean The active nature of the tectonics in the Mediterranean is revealed by the combination of raised (and drowned) shorelines (Chapter 13). For example, when sea levels

Impact on landforms Quaternary volcanic and seismic activity is focused around the migrating arcs. Steep basin and range relief formed by the migrating arcs is associated with the development of a range of landform features. These include (1) alluvial fans associated mainly with mountain range fronts; (2) badlands developed in erodible materials in basin and range areas, particularly tectonic ‘damage’ zones; and (3) subaerial and submarine mass movements triggered by oversteepening of slopes (e.g. by tectonically driven river incision/delta deposition) and seismic activity. Messinian Salinity Crisis develops with a 1 km drop in Mediterranean base level leading to widespread subaerial erosion expressed as deep incision in river systems. Badlands are developed in weak lithologies such as marls and deep karst is developed in limestones. Incision is enhanced by hydro-isostatic uplift around the margins of the Mediterranean. The latter is also associated with drainage re-routing through a combination of defeat of drainages and river capture. Thick evaporite deposits accumulate in the main Mediterranean basin. Reflooding of the Mediterranean causes drowning and sedimentation of karst systems and in the lower reaches of drainage systems. Sedimentary basins satellite to the Mediterranean undergo periodic reflooding and associated evaporite accumulation as they return to normal conditions. The Provençal basin, Valencia trough, and North Algerian basin are opened in the western Mediterranean. The Pannonian back-arc basin is formed as the Carpathians migrate eastwards. The connection of the Tethys Ocean with the Indian Ocean is severed.

Apennines and Carpathian mountains start to develop. The Alps–Betics and Dinarides, Hellenides, and Taurides mountain belts develop.

Development of marine gateways.

were some 3 m higher around 120,000–130,000 years ago during the Last Interglacial, Sardinia preserves their record at +3 m (no change), but the same marine level is elevated to +10 m in Almería, Spain, and south-west Italy, and in parts of Calabria it is more than 120 m above sea level (Ozer and Vita-Finzi 1986 and references therein). This example indicates the marked spatial variations in uplift around the Mediterranean. The temporal changes in uplift rate are also well illustrated from classic sites such as the Gulf of Corinth (Greece). Here some twenty marine terraces have been identified that record relative sea level changes over the last 500,000 years (Keraudren and Sorel 1987) and indicate a regular decrease in uplift rate over that time. Much of the data used to unravel the tectonics of the Mediterranean is based on earthquakes within

8

Anne Mather

the Benioff zone and tomographic images (e.g. Gautier et al. 1999; Lucente et al. 1999; Duermeijer et al. 2000; Wortel and Spakman 2000) and remote sensing data such as Global Positioning System (GPS) and Interferometric Synthetic Aperture Radar (InSAR) (e.g. Kahle et al. 1998; McClusky et al. 2000). The most recent space geodesy confirms an overall north–south compression, with Africa generally moving northwards at 5 mm a−1 (Calais et al. 2003; Nocquet and Calais 2004). It is the north to south convergence of Africa and Europe that has been put forward as the main control on the last 25 Ma evolution of the Mediterranean region. This convergence, a function of differential spreading along the mid-Atlantic ridge, is some 400–500 km in the west and 1,500 km in the east of the Mediterranean. The former Tethys Ocean, generated by the break-up of the supercontinent Pangea, was consumed by this convergence and the Mediterranean Sea was born. Recent work (e.g. Gueguen et al. 1998) suggests that the north–south compression may be secondary in importance to subduction zones migrating orthogonally or obliquely to the main direction of compression, a setting similar to the setting of the modern Caribbean arc. Most recent literature (e.g. Rehault et al. 1985; Malinverno and Ryan 1986; Roydon 1993a; Faccenna et al. 1997; Lonergan and White 1997; Gueguen et al. 1998; Rosenbaum and Lister 2002, Rosenbaum et al. 2002a, b; Rosenbaum and Lister 2004a) favours a model of subduction rollback (Figure 1.2) for the evolution of the Mediterranean, combined with a relatively slow convergence between the two main tectonic plates of Africa and Europe since the Oligocene (Jolivet and Faccenna 2000; Rosenbaum et al. 2002a). Subduction rollback was first applied to the Mediterranean by Le Pichon and Angelier in 1979 to explain the development of the Hellenic arc and trench system in the eastern Mediterranean. Subduction rollback (Elsasser 1971; Molnar and Atwater 1978; Dewey 1980; Royden 1993b) is the result of (1) the negative buoyancy of a subducting slab relative to the asthenosphere (important in the extension of the Tyrrhenian region) and (2) the relative velocity of the subducting slab and the overriding plate or continent (important in the extension of the Aegean region). Differences in buoyancy occur where the subducting slab is cold and dense, such as in old (50 Ma) oceanic slabs (Molnar and Atwater 1978). As a result of the negative buoyancy the subducted lithosphere sinks vertically beneath the asthenosphere, leading to a progressive ‘rollback’ (regressive migration) of the subduction hinge. As the rollback occurs and the hinge migrates, back-arc extension can occur in the overriding

Fig. 1.2. Simplified cross-section of subduction rollback. In (a) V1 and V2 represent the relative plate velocities; P and R represent components of the vertical (negative) buoyancy (F). Where V1 is more than V2 and the subducting slab is old and dense so that R cannot be supported, rollback can occur (b). In (b) back-arc extension is occurring as the rate of subduction rollback (Vr ) is greater than the rate of convergence (Vc ). Modified from Lonergan and White (1997), Rosenbaum et al. (2002a), and Leeder et al. (2003). See text for further explanation.

lithosphere (Royden 1993b) generating sedimentary basins (Figure 1.2). Subduction rollback is still active in the central Mediterranean (e.g Rosenbaum and Lister 2004b) and has been used to explain aspects of the morphology of the Apennine Mountains of Italy (Salustri Galli et al. 2002). Active rollback in the eastern Mediterranean is mostly attributable to the relative velocity of the subducting slab and the overlying plate or continent (Le Pichon and Angelier 1979; Leeder et al. 2003). In the Aegean, for example, the African plate is moving northeastwards at 6 mm a−1 and the Anatolian (Turkish) plate southwestwards at 30 mm a−1 (Briole et al. 2000). Thus the Hellenic subduction zone has to migrate away from both Europe and Turkey as long as these plate velocities are higher. For simplicity, the geodynamics of the Mediterranean presented below will be expressed in the context of the subduction rollback model (Table 1.1 and Figures 1.2 and 1.3). This section will concentrate mainly on the latest publications which synthesize some of the previous work and provide an overall view of the regional geodynamics. A summary of these models can be found in Carminati and Doglioni (2004). Within the present day geodynamic framework there are two subduction zones which are actively destroying

Tectonic Setting and Landscape Development

oceanic lithosphere. These are the Calabrian and Hellenic arcs. There are also a number of collisional mountain belts. Associated with these mountain ranges are a number of back-arc extensional basins which include the Valencia, Provençal, Alboran, Algerian, and Tyrrhenian basins of the western Mediterranean Sea and the Aegean of the eastern Mediterranean Sea. Geologically the Mediterranean can be divided into three sectors—the western, central, and eastern Mediterranean basins (Figure 1.1b). Of these sectors the western

9

Mediterranean is the youngest (mostly less than 30 Ma), and the eastern Mediterranean is the oldest, comprising relics of the Tethys Ocean.

The Western Mediterranean The opening of the western and central Mediterranean (Figure 1.1b) took place over the last 30 Ma by two main episodes of trench migration and back arc opening (Figure 1.3a), which consumed the western Tethys

Fig. 1.3. The present geodynamic framework of the Mediterranean. (a) Map showing the rollback subduction affecting the Calabrian and Hellenic arcs over the last 30 Ma. The arrows indicate the main migration directions of the principal subduction zones over the last 45 Ma or so. (b) A transect across the modern Mediterranean. The location of the transect is shown on the inset map. Modified from Carminanti and Doglionini (2004).

10

Anne Mather

and created the small oceanic basins observable today. The geodynamic evolution of the Gibraltar arc is still controversial. Models proposed for the development of the Betic–Rif mountain belts of Spain and Morocco during the later Mesozoic to Tertiary include (amongst others): (1) westward subduction zone roll-back (e.g. Roydon 1993a; Lonergan and White 1997; Rosenbaum et al. 2002a) and (2) radial extensional collapse of thickened lithosphere (e.g. Platt and Vissers 1989; Houseman 1996). The western Mediterranean is characterized by a large variation in lithospheric and crustal thickness (Figure 1.3b). Seismic and gravity data suggest the lithosphere is less than 60 km thick in the basins (Valencia trough, eastern Alboran Sea, Tyrrhenian Sea) but more than 60–80 km thick in the topographically higher areas such as the Balearic region and Corsica–Sardinia (Carminati and Doglioni 2004). These variations in thickness and composition relate to the Late Oligocene-Early Miocene rifting (Alboran, Valencia, and Provençal basins), which occurred progressively more recently eastwards (eastern Balearic and Algerian basins to the Tyrrhenian Sea), and is attributed to back-arc stretching associated with a westwardsdirected rollback subduction zone which ran east– west along the Maghrebides of North Africa before sweeping northward through the Apennines of Italy (Figures 1.1b and 1.3a; Carminanti and Doglionini 2004). The rollback ensured that the location of the subduction zone migrated eastwards through time (Figure 1.3a), some 700 km in the last 23 Ma. This contrasts to the fact that the north–south relative motion between Africa and Europe at the longitude of Tunisia was only some 135 km over the same time period (Gueguen et al. 1998). This subduction zone consumed the relics of the Tethyan Sea and spread relics of the Alps–Betics orogen located along the subduction zone into the western Mediterranean (e.g. shifting tectonic geography in northern Algeria and Calabria in southern Italy). Rosenbaum and Lister (2004a) estimate that since the Oligocene some 775,000 km2 of the ‘Liguride-Ocean lithosphere’ to the west and some 370,000 km2 of the ‘Adriatic/Ionian lithosphere’ to the east of the western Mediterranean sector have been consumed by the subduction. The pre-existing Alps and Dinarides were boudinaged (extended in sausage-like shapes) during this deformation (Gueguen et al. 1997). The Apennines of Italy formed as an accretionary prism sequence and thus also contain relics of the Alpine orogen. The western Mediterranean as we know it today developed mainly after the final convergence of the Pyrenees at around 20 Ma.

The Central Mediterranean The central Mediterranean preserves a remnant of the Tethys Ocean (the Ionian Sea) between two relict passive continental margins—the Malta escarpment in the south-west and the Apulian (Puglian) escarpment in the north-east (Figure 1.1b). The Sicily channel and the Pelagian shelf between Sicily and eastern Tunisia have been undergoing extension since the Pliocene, demonstrating that Africa is moving south-west in relation to Sicily, generating the grabens of Pantelleria (between Sicily and Tunisia) and Malta. This has deepened the sea floor and generated active alkaline magmatism such as the temporary island of Ferdinandea, to the south of Sicily. It has been suggested (Carminati and Doglioni 2004) that the rifting of the Sicily channel may be linked to rifting observed in south-west Sardinia and the Sirte basin off Libya and through transfer zones through Egypt to the Red Sea and East African Rift system.

The Eastern Mediterranean The mountains of the Dinarides, Hellenides, and Taurides (Figure 1.1b) were generated by the coalescence of a minimum of 2–3 subduction zones since the Mesozoic. The Rhodope–Serbo–Macedonian and Sakayra (northern Turkey) massifs represent the continental margin of the hanging-wall (upper) plate. The Pelagonian (Macedonia–Greece) and Menderes (northern Turkey) massifs represent the continental lithosphere of the footwall (lower plate). These massifs were affected by both compressional and extensional events, but dominated by the latter. This in part explains the lower relief of the Dinarides–Hellenides–Taurides in contrast to other mountain belts such as the Alps and Zagros mountains which have been dominated by compression. North-east directed subduction is still ongoing in the Ionian Sea beneath the Mediterranean ridge (accretionary prism), and beneath Cyprus (Figure 1.1b), which exposes a world-famous geological slice through a complete oceanic section. Convergence rates are at a maximum in the region of the Aegean arc, with plate velocities in the central and southern Aegean reported at 30 mm a−1 (±1 mm a−1 ) in a south-west direction from Global Positioning System (GPS) measurements of crustal motions for the period 1988–97 (McClusky et al. 2000). Since the Miocene, active subduction has also developed along the northern margin of the Black Sea, creating the Caucasus Mountains (Figure 1.1b). The dominant geodynamic feature of the eastern Mediterranean is the north-east subduction of Africa beneath Eurasia. The Aegean Sea is an extensional

Tectonic Setting and Landscape Development

back-arc basin relating to this subduction (Figures 1.1b and 1.3). It has formed as a result of differential convergence rates (Greece is moving south-east faster than Cyprus–Anatolia, relative to Africa; see Briole et al. 2000 and McClusky et al. 2000 for plate vectors and velocities calculated from GPS data) associated with the north-east subduction of the African plate in relation to the hanging-wall Eurasian plate. This zone of extension migrated south-west with time (Figure 1.3a), creating the Aegean Sea, but also affecting continental Greece, Turkey, Bulgaria, Albania, Macedonia, Serbia, and Bosnia. Coevally, the Pannonian basin developed as a back-arc basin of the Carpathians subduction (Figure 1.1b), but migrated eastwards, affecting Austria, Slovenia, Croatia, Hungary, and Romania. In central areas (mainly the former Yugoslavia) it is thus possible to find rifts with opposite directions of migration.

Geodynamics and the Mountain Ranges of the Mediterranean There are a number of mountain belts within the Mediterranean, most of which are attributable to the main subduction zones (Figure 1.1b). What is characteristic of the Mediterranean, however, is the abundance of arcuate mountain belts (Rosenbaum and Lister 2004b). Using the research of Carey (1955) and Marshak (1988) the arcuate mountain belts of the Mediterranean can be classified as ‘rotational arcs’ or ‘oroclines’ (orogens that underwent synorogenic bending) and ‘non-rotational arcs’ (ones that were initiated in their present form). The Jura Mountains of the north-west Alps represent the first case of rotational arcs (oroclines) associated with thin skinned (i.e. thrusting constrained within the sedimentary cover) tectonics (Rosenbaum and Lister 2004b). Other mountain arcs within the Mediterranean seem to have a deeper tectonic origin. These include the Betic– Rif of Spain and Morocco and the Apennine–Calabrian– Maghrebide of Italy and North Africa. Rosenbaum and Lister (2004b) argue that allocthonous fragments of the original mountain belt (e.g. Alpine Corsica, Calabria, Internal Betics) were rotated and moved as independent units before becoming accreted in an arcuate form to the continental palaeomargins of Africa, Iberia, and Adria (central Mediterranean). Work by Krijgsman and Garcés (2004) confirms a lack of rotation about the vertical axis since the Tortonian (which spans 11–17 Ma) in the Gibraltar arc. These data support subduction rollback in a direction that is oblique/orthogonal (i.e. westwards) to the main direction of convergence

11

(northwards) between Africa and Eurasia as the most likely origin for the Betic–Rif arcuate mountain belt, as proposed by Lonergon and White (1997). The significance of this model is that much of the High Pressure/Low Temperature metamorphic rocks in the western Mediterranean that were formerly ascribed to the continental collision of the European and African plates, are in fact the remnants of the extensional destruction of an older (Cretaceous to Oligocene), NE–SW more linear Alpine belt. This belt collapsed as back-arc extension developed (32–30 Ma), and was then dispersed towards the directions of subduction rollback. This led to the subsequent accretion of remnants of that belt during the Oligocene to present subduction rollback described above. At different stages during the Miocene the western subduction systems collided with the palaeomargins of Africa, Iberia, and Adria (Adriatic Sea area). These collisions incorporated continental crust into the subduction system. This decreased the negative buoyancy of the subducting slab and restricted and eventually terminated further subduction rollback. The allochthonous terrains eventually accreted into the continental palaeomargins, generating the arcuate mountain belts. Although some of the overall curvature of the mountain belts can be explained by bending of the lithospheric slabs, to account for the tightness of arcuate features such as the Betic–Rif belt implies that some tearing of the subducting slabs must have occurred (Rosenbaum and Lister 2004b).

Distribution of Modern Seismic Activity Seismic activity within the Mediterranean is geographically diffuse. Although earthquake magnitudes are typically not large, the density of population means that they do provide a significant hazard (Chapter 16). Some of the highest recorded magnitude earthquakes within the Mediterranean are reported from North Africa (6.9 Mw in 1980); Sicily (7.5 Ms in 1908) (Ms is earthquake magnitude calculated from surface waves), and Turkey (7.6 Mw in 1999 and 8.0 Mw in 1939) as documented in Vannucci et al. (2004) and references therein. Data from Vannucci et al. (2004) indicates that most shallow earthquakes (<50 km deep) are associated with the main plate boundaries (Figure 1.4a: Africa, Eurasia, and Arabia), but other clusters hint at the presence of a number of other micro plates in the region. The most concentrated seismic activity

12

Anne Mather

Fig. 1.4. Seismic activity in the Mediterranean. (a) The distribution of shallow earthquakes (<50 km depth) across the Mediterranean (modified from Vannucci et al. 2004). (b) The distribution of earthquakes occurring at >50 km depth (modified from Vannucci et al. 2004) and Holocene volcanic activity (data from the Smithsonian National Museum of Natural History Global Volcanism Project, available online at , accessed 29 Sept. 2008).

is around the Hellenic arc, the Aegean Sea, and the Zagros Mountains (Figure 1.4a). The main belt of seismicity can be traced from the Mid Atlantic Ridge in the west, dominantly through the main mountain ranges of the Mediterranean. The zone passes through the Gibraltar arc, Northern Africa to Sicily, the Tyrrhenian basin, Calabrian arc and Apennines, into the Alps then the Dinarides–Helenides, Hellenic arc, and Zagros (Figure 1.4a). Deeper earthquakes (>50 km deep,

Figure 1.4b) tend to be concentrated in those areas associated with lithospheric subduction (Chapter 16). In the Hellenic arc earthquakes occur down to a depth of 200 km, although the subducted slabs go much deeper to some 670 km (Lucente et al. 1999; Wortel and Spakman 2000), beneath Calabria. There is some controversy, however, as to whether this slab is actively subducting or detached from the surface (see references in Rosenbaum and Lister 2004a ). In the

Tectonic Setting and Landscape Development

southern Tyrrhenian Sea earthquakes occur down to 600 km, associated with an approximately 200 km long Benioff–Wadati Zone (Rosenbaum and Lister 2004a). Southern Spain and the Carpathians are associated with earthquakes of intermediate depths (60–120 km) of large but sporadic occurrence, and some earthquakes down to 600 km (Buforn et al. 1991). Some of the intermediate-depth earthquakes below Gibraltar may correspond to active eastward subduction in this area (Gutscher 2002) although there is no present-day magmatic activity to support this (Figure 1.4b).

Distribution of Modern Volcanic Activity Modern volcanic activity within the Mediterranean is mainly constrained to the areas of the Calabrian and Aegean arcs although volcanism active over the Holocene and Pleistocene covered a much wider area (Figures 1.4b and 1.5). The volcanic region which runs from Tuscany to Sicily and Pantelleria is the largest volcanic region in Europe and activity began here in

13

the Pliocene. Active volcanoes include Etna, Stromboli and Vulcano, Vesuvius, the Solfatara of Pozzuoli, the Phlegraean Field, and the Island of Ischia (Chapter 15). The Aegean island arcs include the active volcanoes of Methana, Santorini, and Nisyros. These have been generated from subduction along the Hellenic trench and partial melting of the descending slab from about 150 km depth which has generated calc-alkaline (largely dacitic) magmatism. The activity here is continuous, but strongly affected by tectonic reorganization of the plates. The latter has been documented at 0.8– 0.5 Ma, when rapid trench migration and Tyrrhenian back-arc extension stopped. The net result of this was that the regional plate movements were then accommodated by a transfer to a back thrust in the South Tyrrhenian Sea, which has been the source of frequent M5 to M6 earthquakes over the last 20 years (Goes et al. 2004). A large but diffuse transform boundary fault developed across north-east Sicily, connecting with Calabria along the line of the Aeolian Islands (Salina, Lipari, and Vulcano) and opening the Messinia straits between Sicily and Calabria. The latter led to a period of enhanced seismic activity and unusual volcanism (carrying an

Fig. 1.5. Volcanic activity in the western Mediterranean over the last 33 Ma. Adapted from Rosenbaum et al. 2002a. Note how the activity is older (15–7 Ma) in the western, Alboran Sea, reflecting an older, east dipping subduction zone that retreated westwards during the Middle–Late Miocene (Lonergan and White 1997). Since the Late Miocene most volcanism has been concentrated in the Tyrrhenian Sea where a west dipping, eastward migrating subduction zone is located. The most recent activity is in the Calabrian arc (Chapter 15).

14

Anne Mather

ocean-island basalt geochemical signature) in the Aeolian Islands and Mount Etna (Goes et al. 2004). Stromboli is the most continuously active of the Aeolian Island volcanoes. The presence of the largest active volcano in Europe, Mt. Etna (3,350 m a.s.l.), is not so easily explained. This exists on the hanging wall of the accretionary wedge and is produced by Na-alkaline magma. Doglioni et al. (2001) suggest that this may relate to right lateral transfer along the Malta escarpment producing a transtensional ‘window’ between the Sicilian and Ionian components of the Apennines slab. This was thought to be generated by differential rollback enabling Etna magma to be sourced from the lower lithosphere to upper asthensophere.

Palaeo-oceanography of the Mediterranean Basin and Changing Marine Gateways The modern Mediterranean is a semi-enclosed basin that links with the main oceans through the Atlantic via the Straits of Gibraltar. This gateway is fundamental in controlling the maritime conditions within the Mediterranean Sea (Chapter 2). Formerly, the number of gateways was greater but convergence of the African and Eurasian plates led to the progressive closure of these links. In the east, the Tethys Ocean (forerunner to the Mediterranean Sea) had its connection with the Indian Ocean severed in the Middle Miocene (Hsü and Bernoulli 1978). The loss of this marine gateway dramatically affected ocean circulation patterns and was thought to be responsible for a cooler global climate (Miller et al. 1991 and Chapter 4). Ever since this event the Mediterranean Sea has been prone to a negative water balance. In the west two marine gateways used to exist. These were the Betic Corridor, southern Spain, and the Rifian Corridor, northern Morocco. These were progressively closed during the Late Miocene due to an interaction between tectonic and glacio-eustatic processes (Krijgsman et al. 1999a ; Garcés et al. 2001). Their closure is associated with the Mediterranean Salinity Crisis (see below). The Betic Corridor is believed to have closed in the Tortonian leading to a ‘Tortonian salinity crisis,’ which predates the better-documented Messinian Salinity Crisis (Krijgsman et al. 1999a; Soria et al. 1999; Krijgsman et al. 2000). The Rifian corridor was the deepest (some 600–800 m; Hodell et al. 1989; Krijgsman et al. 1999a) and was emergent at 6.1 Ma and thus also predates the Messinian Salinity Crisis (Krigsman et al. 1999a; Garcés et al. 2001). The

modern gateway between the Mediterranean and Atlantic is thought to have formed during the Pliocene (Hsü et al. 1973) and led to the reconnection of a Mediterranean–Atlantic gateway and a return to more normal marine conditions within the Mediterranean basin.

The Messinian Salinity Crisis and its Geomorphic Legacy The Messinian Salinity Crisis occurred when the last of the ocean gateways closed and restricted any connection with the Atlantic Ocean. This led to an evaporitic drawdown of the Mediterranean, which led to widespread precipitation of evaporites (Hsü et al. 1973) within the Mediterranean basin and satellite basins (e.g. the Sorbas basin, south-east Spain). The storage of salts from the evaporated sea water is calculated to have reduced the world ocean salinity levels by some 6 per cent and temporarily modified global thermohaline circulation (Hsü et al. 1977). This drop in salinity would have increased the extent of sea ice and may have contributed to the onset of the next glacial phase. The age of the event has been strongly debated. Some workers (e.g. Baumard 2001) suggest the Messinian Salinity Crisis spanned 5.8–5.32 Ma. On the basis of astrochronology Krijgsman et al. (1999b) suggest that the precipitation of evaporites began at 5.96 ± 0.02 Ma and that complete isolation from the Atlantic was achieved between 5.59 and 5.33 Ma. The evaporitic drawdown caused a dramatic fall in the Mediterranean Sea level (>1 km) and the development of extensive on-shore erosion (5.59–5.50 Ma) and deposition (5.50–5.33 Ma) of non-marine sediments in a lake sea (‘Lago Mare’). Krijgsman et al. (1999b) report that the event terminated with reflooding around 5.62–5.50 Ma and that the Mediterranean returned to normal marine conditions in the Pliocene around 5.33 Ma.

Causes It is generally agreed that glacio-eustatic fall (Adams et al. 1977; Hodell et al. 1986), horizontal crustal shortening (Weijermars 1988), tectonic uplift (Garcés et al. 1998; Hodell et al. 1989; Krijgsman et al. 1999a ) and climate change were the most likely triggers for the onset of the salinity crisis. Taking each of these mechanisms in turn we can attempt to ascertain their relative significance. With regard to a glacio-eustatic fall, the proposed 60 m drop in sea level in the Messinian is insufficient to have closed all the marine gateways. Also, the onset of evaporitic deposition at 05.96 ± 0.02 Ma

Tectonic Setting and Landscape Development

does not correspond with the benthic glacio-eustatic ‰18 O signature, suggesting that the Messinian Salinity Crisis is unlikely to be related to glacio-eustasy (Krijgsman et al. 1999b; Hodell et al. 2001). Palynological work in the Rifian corridor of northern Morocco (Warny et al. 2003) suggests that the regional climate was stable before, during, and after the Messinian Salinity Crisis. Crustal shortening and nappe emplacement is unlikely as it predates the Messinian Salinity Crisis (Comas et al. 1999). Thus the favoured mechanism is uplift. Duggen et al. (2003) account for the necessary uplift of c.1 km through the westward rollback of the subducted Tethys Oceanic lithosphere. With regard to the reflooding of the Mediterranean, data suggest that despite the Rifian corridor being the deepest of the oceanic gateways, reflooding probably occurred through the Gibraltar arc (Warny et al. 2003), creating a new marine gateway—the Strait of Gibraltar. The location of this gateway has been attributed to a mantle origin—involving gravity-induced slumping and faulting—by some workers (Duggen et al. 2003), perhaps assisted by geomorphic processes such as stream piracy (Blanc 2002). In the latter scenario an eastwardflowing stream that drained the eastern slope of an emergent Gibraltar Isthmus is thought to have breached the Atlantic/Mediterranean watershed. Blanc (2002) modelled the Atlantic inflow and suggests that the rate of Mediterranean recharge was exponential, with the level of the Mediterranean basin hardly changing in the first 26 years, but that it was completed within the ensuing 10–11 years. These dramatic and (on a geological timescale) extremely rapid changes in base level had profound consequences for the Messinian geomorphology of the region. The impacts are still visible in the regional Quaternary geomorphology.

Geomorphic Legacy The main geomorphic legacy of the Messinian Salinity Crisis relates to: (1) the dramatic changes in base level for the Mediterranean region; (2) the hydro- and erosional isostatic readjustments and (3) the widespread accumulation of evaporitic deposits and their associated distinctive karstic terrains. The dramatic lowering of base levels is credited with developing Mediterranean basin-wide subaerial erosion surfaces (Ryan 1978; Ryan and Cita 1978). The lowering of the Mediterranean Sea during the Messinian Salinity Crisis has been estimated at 500– 1,000 m (Mauffret 1979; Durand Delga 1980), 1500 m (Schlupp et al. 2001), 1,900–2,400 m (Gargani 2004), 2,500 m (Le Pichon et al. 1971; Ryan and Cita 1978;

15

Clauzon 1982) and 3,000 m (Malinerverno et al. 1981). However, the identification of the major erosion surface in some of the basins satellite to the main Mediterranean is hotly debated (see e.g. Fortuin et al. 2000 and Riding et al. 2000 and references therein). This erosion would have locally exacerbated any hydro-isotatic uplift as a result of erosional unloading (e.g. Gargani 2004). The presence of entrenched meanders and terraces within the Nile Canyon (formed when the Proto-Nile cut into the underlying Tortonian delta sediments as a result of the Messinian base-level fall) suggests sudden sporadic falls in sea level rather than a progressive drop (Barber 1981). Similarly, numerically modelled data from the Rhône suggests an initial drop in sea level of around 600–700 m followed by a secondary drop of 1,300–1,700 m. Gargani (2004) proposes that the first drop occurred over a period of 400 ka, and the second over a period of 50 ka. The associated loading and unloading of the Mediterranean basin water led to significant hydro-isostatic rebound. Two-dimensional flexure models (Norman and Chase 1983) suggest that the north-western and southeastern coasts of the Mediterranean would have generated shoreline bulges some 450 m high that were capable of reversing rivers with low stream power. Only those rivers with sufficient stream power (e.g. the Ebro, Po, Rhône, and Nile) could maintain their original courses through some of the areas of peripheral bulge. The hydro-isostasy would have been emphasized in the deepest parts of the Mediterranean basin such as the Ionian, Provençal, and central Tyrrhenian seas (Carminati and Doglioni 2004).

Mediterranean Rivers The main rivers of the Mediterranean, such as the lower Proto-Rhône and Ebro and Nile rivers, exhibit kmscale incision as a result of the dramatic lowering of base level in the Mediterranean during the Messinian (Clauzon et al. 1996). Outcrops and seismic data indicate deep palaeovalleys infilled with Pliocene sediments in the Rhône and Durance valleys (Audra et al. 2004). Canyon incision in the Aegean has been attributed to capturing the Black Sea drainage and leading to the deposition of ‘Largo Mare’ sediments (Hsü et al. 1973; McCulloch and De Deckker 1989). In response to the drop in base level a wave of incision up the main valley systems lead to headward erosion of valleys tributary to rivers such as the Po Canyon. These tributaries head-cut back into the Alpine highlands, forming the valleys later exploited by Quaternary glaciations. The use of boreholes and seismic mapping of the Nile Delta region (Barber 1981) shows that the basal

16

Anne Mather

Messinian erosion surface extends from approximately sea level in the southern delta area to depths of 5 km in the Nile cone. In the Nile region there is also subsurface evidence of the Messinian erosion surface that developed, and its associated drainage patterns. These appear to indicate that the lithology exerted a strong control on the development of the drainage pattern, with trellis drainage developed on resistant Oligocene basalts and dendritic drainage on less resistant Tortonian prodelta shales. The drop in base level of the Mediterranean led to river capture of the trellis drainage by the dendritic networks (Barber 1981) leading to sediment re-routing along parts of the Mediterranean shoreline.

Mediterranean Karst Karst landscapes within the Mediterranean region are commonly developed in Cretaceous limestones (Chapter 10), and more uniquely in the Messinian gypsum relating to the Messinian Salinity Crisis, and thus postdating it (e.g. Calaforra and Pulido-Bosch 2003; Ferrarese et al. 2003). The karsts developed within Cretaceous limestone often show a complex history that can be related back to the Mediterranean Salinity Crisis (Audra et al. 2004). Deep karst systems were probably formed during the Messinian Mediterranean drawdown, and reflooded during the Pliocene when the Mediterranean salinity crisis terminated. It is possible that palaeokarst systems such as the Rospo Mare oilfield (in Cretaceous limestone) located some 1,200 m below the modern Mediterranean sea level in the Adriatic (Soudet et al. 1994) may also owe their origin to this phase of the Mediterranean’s history. Off the coast of France the submarine spring of Port-Miou, south of Marseilles, is located in a drowned canyon of the Calanques Massif (Audra et al. 2004). Here the main water flow comes from a vertical shaft >147 m below present sea level. The shelf margin comprises a submarine karst plateau cut by a deep canyon, with the canyon reaching 1,000 m below sea level. In southern France inland karst systems are unusually deep, reaching >224 m below sea level at Fontaine de Vaucluse (Chapter 10). At the latter site flutes on the side of the vertical conduit, observed with a remote observation vehicle, indicate that this shaft was once air filled and connected to an underground river that was flowing towards a deep valley (Audra et al. 2004). Within the Ardeche a number of Vauclusian springs probably relate to the Messinian Rhône Canyon which is located 200 m below present sea level. Similarly the formation of karst systems in the Languedoc of southern France has also been linked with the drawdown of the Mediterranean during the salinity crisis (Josnin 2001). Josnin (2001) also suggests that

one of the primary controls on the karst in this region is the fracturing developed during the major tectonic phases of the Mediterranean. Within the latter systems reactivation of faults within the karst has led to the unblocking of abandoned conduits and their reincorporation into the active system. Mediterranean karst environments are discussed in detail in Chapter 10.

Geodynamics and the Nature of Quaternary Landforms It has already been shown that a combination of north– south compression and subduction rollback oblique/ orthogonal to this can explain the distribution of the main mountain ranges and basins of the Mediterranean region on geological (106 year) timescales. These mechanisms have provided the template for the landscape of the modern Mediterranean and can be used to explain the distribution and nature of key landforms.

Mountain Range Morphology The morphology of mountain ranges is largely controlled by tectonics, eustasy, climate, and lithology (e.g. Allen 1997; Leeder et al. 2002; Silva et al. 2003). Together these govern rates of erosion that, balanced against areas of crustal uplift, will lead to land elevation where erosion is less than uplift, i.e. surface uplift is occurring. Perhaps the most studied mountain range in the Mediterranean in this context are the Apennines of Italy. These are a fold and thrust belt that forms an accretionary wedge above the westward-directed Appenine– Maghrebides subduction zone (Lenci et al. 2004). Within mountain ranges the hydrological drainage divide and the highest elevations in mountain belts usually coincide, but they may move or grow as a function of the vergence and rate of the tectonic evolution (e.g. Ollier 1995). In the Appenines of Italy the most elevated topography is located eastward of the drainage divide. This has been attributed to lateral (eastward) migration of the topography associated with plate tectonic slab rollback (Figure 1.6). Lenci et al. (2004) used seismic lines to examine the relief and cross-sectional area of Italian mountain ranges. They found that in the Apennines where subduction depth was 200 km, and décollement depth (the depth of detachment of the upper cover from its substratum) was 10 km, the cross-sectional area of the mountains was 2000 km2 and relief greatest. Conversely, in the more southerly Calabrian arc, the subduction depth was 500 km, décollement depth was 3 km, the cross-sectional area of the mountains was

Tectonic Setting and Landscape Development

17

Denudation rate <1 mm/a Migrating system 10 – 30 mm/a

W

E Extensional belt

t1

3

Accretionary wedge

t2

t3

t4

2 km 1 0 0

50 km

Vertical exaggeration x 10

Slab rollback 10 – 30 mm/a

t1

t2

t3

t4

Fig. 1.6. Eastward migration of the topography in conjunction with the eastward rollback of the subduction zone in the Apennines of Italy through time (t1–t4). Modified from Salustri Galli et al. (2002). The simplicity of this model will be modified with variations in rollback migration, lithological variations in the accretionary wedge, and the depth of décollement.

1,500 km2 and the overall topographic relief was less. Thus the décollement depth appears to control the volume of material that becomes involved within the accretionary prism, and thus the topography of the mountain range.

Mediterranean Rivers As a result of the regional tectonics of the Mediterranean, most of the associated river systems (apart from the exogenous River Nile) tend to be constrained by the uplifted mountain ranges and are thus small and proximal to the Mediterranean Sea. The largest drainage areas are associated with the Ebro (84,230 km2 ), the Po (70,090 km2 ), and the Rhône (95,590 km2 ) (Macklin et al. 1995). The river systems can be grouped into (1) steepland rivers, typically above 500 m elevation and

(2) basin and range drainages (Macklin et al. 1995). The first of these tend to have high stream power as a result of steep gradients and carry high sediment loads, often sourced from mass movement events. In the more arid parts of the Mediterranean (e.g. south-east Spain and the Near East) these upland streams tend to be ephemeral and associated with low-frequency, highmagnitude flood events, whilst in the northern Mediterranean they may experience high-magnitude floods in association with snowmelt (Chapter 8). A few of these drainage systems are developed on active volcanoes such as Mount Etna in Sicily. These are subjected to aggradation linked to high sediment production during volcanic events (e.g. Chester and Duncan 1982). The second drainage configuration associated with the basin and ranges tends to comprise incising bedrock channels in uplifted fault blocks and adjacent sediment

18

Anne Mather

receiving areas in the downthrown block (graben). The grabens are often associated with axial-draining larger alluvial systems that may be braided or meandering (e.g. the classic Büyük Menderes River of western Turkey). In these regions the drainages can be segmented and diverted by active faulting and tilting (see, for example, Goldsworthy and Jackson 2000) and the drainage basin length scales and areas are related to the spacing and geometry of the major normal faults, together with the susceptibility of the exposed lithologies to erosion (Collier et al. 1995). The fluvial systems may be associated with aggradational features such as alluvial fans at mountain fronts (see below) and within tributary valley junction, or fan deltas at the coast. The main delta areas in the Mediterranean are the Nile (22,000 km2 ), the Po (770 km2 ), the Rhône (720 km2 ), and the Ebro (350 km2 ) (Macklin et al. 1995 and Chapter 13). These latter areas of sediment build up are often associated with significant submarine mass movements as a function of sediment loading and seismic triggering. Detailed work by Amorosi et al. (1996) links the incisional records of upland rivers with alluvial deposition at the coast. They recorded up to 32 terrace levels from the Reno River of the northern Apennine mountains. These upland catchments are located within an active thrust belt. They have supplied more than 300 m thickness of alluvium to the receiving Po basin (the adjacent foreland basin) over the last 35 ka. These authors identified that deep valley incision, which is associated with poor river terrace preservation (as a result of high erosion rates) in the Apennine mountains could be correlated with enhanced alluvial deposition where the rivers enter the receiving Po basin. These enhanced phases of geomorphological activity in the region occurred in response to rejuvenation of relief (i.e. uplift generated relative drops in base level) during successive thrusting events.

Alluvial Fan Distribution Alluvial fans occur throughout the Mediterranean region. Although tectonics is not a prerequisite for fan formation, it is responsible for providing suitable topographic locations for the fans by generating the high relief in the Mediterranean region essential for their formation. Thus many Quaternary fan systems are located at mountain fronts (Figure 1.7) and have thus been used to elucidate palaeoseismicity of mountain front faults (e.g. Michetti et al. 1997; Martinez-Diaz et al. 2003; Galadini and Galli 2003). Although the depositional sequences and morphology may be modified by local tectonics (e.g. Silva et al. 1992; Orri 1993) and sea level

Fig. 1.7. An example of alluvial fans from a faulted mountain front in the Tabernas basin of south-east Spain. The smaller, debris flow dominated fans (one is outlined) are less than 1 km long and 0.5 km wide, with an area of 24 ha. The larger, fluvial dominated fan is 2.8 km long, 1.7 km wide and some 374 ha in area. (Image courtesy of the Junta de Andalucía Ortofotografía digital de Andalucía, ISBN 84-95083-96-5.)

(Harvey et al. 1999), most fan sequences themselves are climatically driven (e.g. Harvey et al. 2005).

Badland Distribution Badlands are areas of extensive gullying (first named ‘Mauvaises Terres’ by French settlers encountering the North Dakota badlands of the USA in the eighteenth century; Grove and Rackham 2001). Badlands are distributed across the Mediterranean in sparsely vegetated areas (e.g. the Tabernas basin of south-east Spain, Figure 1.8) to ‘partial’ areas of badlands in wooded areas (e.g. the Aegean island of Rhodes) and thus span both the drier and wetter areas of the Mediterranean climate (Table 1.2). Much badland development has been attributed to human impact (land mismanagement) but there is good evidence to support the fact that the origin of most badlands predate major human impact within the region and are typically controlled by longer-term controls such as regional tectonics. For example, present measurements of erosion in the Guadix badlands of southern Spain are very low (0.01 mm a−1 ) and the preservation of many Copper and Bronze Age archaeological sites located on ridges and slopes within the badlands indicates stability for some 3,500–5,000 years to account for their preservation (Grove and Rackham 2001). Badlands in southern Israel can be dated back 70 ka (Yair et al. 1982) and even older badlands are attributed to the drawdown of the Mediterranean

Tectonic Setting and Landscape Development

19

Fig. 1.8. An example of well-developed badlands in Tortonian marls within the Tabernas basin of south-east Spain. Note the motorway in the top left background for scale. The modern badlands are inset below pediment surfaces which indicate a former, higher base level. (Photo: Anne Mather.)

Sea and associated widespread erosion in the Messinian (Ryan 1978). Regional-scale tectonic processes in the Mediterranean affect the distribution of badlands by controlling

the primary geological distribution of susceptible materials (those of low unconfined compressive strength such as marls and phyllites) and, through deformation (uplift, faulting etc.), the exposure of these susceptible sediments

TABLE 1.2. Examples of badlands and partial badlands (∗) found across the Mediterranean Source Harris and Vita-Finzi (1968) De Ploey (1974) Alexander (1982) Harvey (1982) Imeson et al. (1982) Yair et al. (1982) Woodward et al. (1992)* Alexander et al. (1994) Prinz et al. (1994) Woodward (1995a, b)* Gallart et al. (2002)*

Mean Annual Rainfall (mm) 1,500 350 450 170–350 300 90 2,000 170 300–50 1,500–2,000 924

Location Kokkinopilos, Louros Valley, NW Greece Kasserine, Central Tunisia Agri basin, Basilicata, Italy Almería to Alicante, SE Spain Rif Mountains, Morocco Northern Negev, Israel Pindus Mountains, Epirus, NW Greece Tabernas, Almería, SE Spain Northern Algeria East of Konitsa, Pindus Mountains, NW Greece Vallcebre Catchment, Pyrenees, Catalonia, Spain

Source: Modified from Campbell (1989) and Woodward (1995) (with additions).

Lithology Red silts and clays Clays, loams, and sandy lithosols developed on marls Clays, silts clays, interbedded sands, soft shales, and mudstones Marls, silts, shales, and sandstones Clays, silts, and sands Marls and soft shales Interbedded sandstones and fissile siltsones Marls and rare interbedded sandstones Marl soils Shales and sandstones Smectite-rich mudstones

20

Anne Mather

to surface erosion. Lithology is a strong primary control on badland distribution and rates of erosion. In the Ebro badlands of north-east Spain Tertiary clays were found to erode at 15–22 mm a−1 (Benito et al. 1992), whereas Holocene sediments (in the same area and study) eroded at a third of this. Badland susceptible materials within the Mediterranean are usually located within the ‘damage zone’ of major tectonic structures and are thus affected by joints, faults, and associated features (e.g. gypsum veining) which will control the hydrology of the slopes and thus gully and pipe alignments (e.g. Harvey 1982). Thus badlands are prolific in tectonically active areas such as the basins of south-east Spain, the Apennines of Italy, and the Corinth fault zone of Greece.

Mass Movement Activity The relief generated by the Mediterranean geodynamics is partly responsible for the abundance of mass movements in the region (e.g. Murphy 1995; Thornes and Alcantara-Ayala 1998). Many of these mass movements are triggered by rainfall events (Chapter 6), particularly in the higher elevation areas (e.g. Corominas and Moya 1999; Calcaterra and Santo 2004; Fiorillo and Wilson 2004; Luino 2005) or individual earthquake events (e.g. Wasowski and Del Gaudio 2000; Keefer 2002; Porfido et al. 2002, Carro et al. 2003;

(a)

Papadopoulos et al. 2003; Reicherter et al. 2003). Historical records, however, tend to greatly underestimate the number of landslides attributable to a given earthquake event (Keefer 2002). Several authors have noted correlations between earthquake magnitude, distance from epicentre, and landslide occurrence (e.g, Keefer 1994; Rodriguez et al. 1999; Papadopoulos and Plessa 2000). Seismically triggered mass movements are known to generate local tsunamis (e.g. Darawcheh et al. 2000; Maramai et al. 2005a , b), adding to the tsunami hazard in the Mediterranean (Chapter 17). For example, on 30 December 2002, seismic activity on Stromboli triggered a submarine slump of 20 million m3 (Pino et al. 2004) and generated a local tsunami several metres high (Figure 1.9). The magnitude of individual subaerial mass movements ranges from small rockfalls to those of sufficient size to dam valleys (34–50 million m3 reported by Nicoletti and Parise 2002 from south-eastern Sicily). However, the largest mass-movement events are submarine in origin (Lastras et al. 2004). These include a 22 ka, 500 km3 mega turbidite in the Balearic Abyssal Plain (Rothwell et al. 1998, 2000); and a 120 km3 debris flow in the eastern levée of the Rhône deep-sea fan (Droz 1983; Gaullier et al. 1998; Méar 1984). Recent earthquake-triggered turbidity currents also occurred off the coast of Algeria (Heezen and Ewing 1955; El-Robrini et al. 1985). The submarine BIG’95 Late glacial–early Holocene slide

(b)

Fig. 1.9. Associated volcanic activity and tsunami impact in the Aeolian Islands. (a) Evidence of activity on Stromboli (image taken 3 January 2003 by Anne Mather). The activity began 30 December 2002 with seismic activity triggering a landslide that generated a tsunami. Image (b) shows the impact of the tsunami wave on Stromboli (image courtesy of Victor Kakebeeke). Note the erosion created by the backwash.

Tectonic Setting and Landscape Development

(discovered by, and named after the cruise ‘Biogeoquimica I Geologia’ in 1995; Canals et al. 2004) was generated on the western side of the Valencia trough which is one of the Oligocene/Early Miocene-Pleistocene extensional basins. The Ebro margin here comprises a 70-km-wide shelf, a 10-km-wide slope, and a smooth continental rise that progressively deepens into the Valencia Channel. The base of the continental slope lies at 1,300–1,800 m below sea level. The BIG’95 landslide buried the upper course of the Valencia Channel. The slide contains gliding blocks and debris flows and covers an area of >2,000 km2 with a volume of 26 km3 . The total run out distance was some 110 km, with individual blocks up to 12,500 m × 3,000 m travelling some 3 km. It is considered that the likely triggering mechanism was a local earthquake event some 11,647–11,129 cal years BP (Canals et al. 2004).

Quaternary Landscape Development: Case Studies The main landform features discussed above will evolve in response to spatial and temporal changes in rates and styles of deformation. In addition the landforms do not act as separate entities but will interact to give an often complex response to ongoing tectonics. The developing landscape may reflect (1) direct tectonic control or (2) indirect tectonic control. Direct responses to tectonics occur in areas of rapid rates of deformation (e.g. the eastern Mediterranean). In the case of river systems this may lead to drainage diversion, for example reversed drainage. In areas of lesser tectonic activity (e.g. the western Mediterranean) the impacts of deformation may be less clear, but may be associated with equally dramatic river re-routing via river capture. Below, two case studies are used to illustrate the direct and indirect landscape responses to tectonics from two detailed case studies from the eastern and western Mediterranean. A recent special issue of Geomorphology edited by Silva et al. (2008) examines the impact of active tectonics on fluvial landscapes and drainage network development. With a strong focus on the Mediterranean, it includes case studies from the High Atlas of Morocco, the Dead Sea Rift, the Apennines of Italy, and various parts of the Iberian Peninsula.

Direct Impact of Tectonics: Southern Greece The direct impact of tectonics on the Quaternary landscape is perhaps best illustrated from the classic Gulf

21

of Corinth areas of southern Greece. Most of Greece is undergoing extension and subsidence as a result of subduction along the Hellenic arc, but areas such as northern Greece, where the Pindus Mountains are located, are undergoing long-term compression and uplift. Overall the landscape is made up of mainly east–west orientated grabens separated by faults from actively uplifting areas. This has created topographic features such as the Gulf of Corinth. Much of this active deformation relates to the motion between small plates around poles of rotation for each of these plates. In the central Mediterranean these plates are the Aegean, Ionian, and European plates and the poles are the Aegean–Ionian pole, the Aegean–European pole, and the Ionian–European pole (Figure 1.10). Of these boundaries the Aegean–Ionian boundary is typified by active subduction which varies from 3 cm a−1 in the north (Levkas) to 6 cm a−1 towards the south (Crete). The Aegean–European boundary is extensional with rates of movement of 2.5 cm a−1 in the west (Gulf of Corinth) to 5 cm a−1 in the east (Turkey). The relative poles of rotation generate these regional variations in rates of movement (Figure 1.10). The resultant tectonic style associated with these plates and their relative movements has uplifted mountainous areas such as Crete and Levkas (Chapter 16). The region is associated with large earthquakes on reverse

Fig. 1.10. Plate boundaries and the main poles of rotation referred to in the text. The poles presented here are those relevant to the motion at the plate boundaries of north-west Greece. The box indicates the location of the Gulf of Corinth and the image in Figure 1.11. Modified from King et al. (1997).

22

Anne Mather

faults in areas of compression and active seismicity and volcanism (Aegean Sea) in areas associated with the steeply (45◦ ) subducting Ionian Sea floor. The extension and compression which has affected the region has led to the development of the largest scale morphological features observable within the region, i.e. the basins and ranges. These features have often been inverted with former sediment receiving basins developed during extension becoming uplifted and exhumed source areas during compression. This section will now focus on one of these areas to examine the impact of these active tectonic processes on the associated geomorphological systems. The selected area is the welldocumented Gulf of Corinth and its eastern extent— namely the Alkyonides Gulf and inverted Megara basin (Figures 1.10 and 1.11).

Over the last 300 ka the uplift on the southern side of the Gulf of Corinth has been documented by the development of a number of erosional marine terraces formed during periods of high sea level (Kerauden and Sorel 1987; Armijo et al. 1996). The cumulative uplift for these terraces has been calculated at 1.5 mm a−1 but decreasing with time (Kerauden and Sorel 1987), and the relative displacement of blocks either side of a fault at 1 cm a−1 (King et al. 1997). These terraces are subject to active erosion and the rapid land degradation from gullying in this region can be clearly linked with active faulting (King et al. 1997). Over shorter timescales (individual earthquakes) significant deformation can occur rapidly. For example, in 1981 an earthquake of Ms 6.7 on 24 February affected the area of Corinth. Aftershocks of Ms 6.4 on 25 February and 4 March did further

Fig. 1.11. Oblique view of the Megara basin created with 3× vertical exaggeration using Geocover 2000 Landsat TM data courtesy of NASA World Wind. Note the drainage to the Alkyonides Gulf at the eastern end of the Gulf of Corinth cuts an erosional cirque into what is the footwall of the active Alepochori fault. Some 9 km3 of sediment has been removed by this drainage (current limit of the drainage divide is shown by the dashed line to the left of the image) over the last 1 Ma (Leeder et al. 1991). The drainage on the backtilted footwall shows tributaries dominantly draining towards the Megara fault, and the main axial drainages draining perpendicular to this trend, into the Saronic Gulf. The dotted line to the right of the image indicates the limit of stream incision on this surface. This asymmetric stream pattern reflects (1) an inherited tilt towards the inactive Megara fault, generating the early, north-east tributary drainage orientations, (2) backtilt from the active Saros fault, maintaining the NE/SW tributary drainage orientations and (3) backtilt from the active coastal scarp fault in the Gulf of Corinth generating the south-east orientated axial drainages.

Tectonic Setting and Landscape Development

damage. These led to significant geomorphological changes with some parts of the coast sinking by 1 m (Jackson et al. 1982; Vita-Finzi and King 1985; King et al. 1997; Chapter 16). Thus it is not uncommon to find juxtaposed areas of uplift and subsidence along some of the major active normal faults, with significant geomorphic implications (e.g. Leeder et al. 1991). The Alkyonides Gulf, at the eastern end of the Gulf of Corinth, is bounded to the south by a series of fault scarps, some of which were active during the 1981 earthquake (Jackson et al. 1982). Some of the individual fault scarps moved by as much as 1.5 m vertically, but averaged 0.5 to 0.6 m (Jackson et al. 1982) in the 1981 events. The topography associated with the main fault system descends to 400 m below sea level and the adjacent mountains are 1,000 m above sea level (Leeder et al. 1991). Indicators of active tectonics (hanging-wall subsidence) include drowned alluvial fans and associated marshes in the Psatha Bay beach zone (Leeder et al. 1991). These marsh areas are being progressively transgressed by backbeach washover sedimentation. The location of many marine (bays, lagoons, beach rock reefs, raised beaches) and continental (debris cones, alluvial fans) features is controlled by faulting and associated seismicity in this region (Leeder et al. 1991). Some of the debris cones, for example, are attributable to the 1981 sequence of earthquake events. Uplift in the footwall of the main Psatha-Skinos fault also controls stream incision, encouraging active headward erosion of stream networks (Leeder et al. 1991; Collier et al. 1993). Rates of drainage expansion and erosion within the area are mainly controlled by lithological variations associated with the faulting. The Megara basin of Greece is an inverted basin located at the eastern end of the Alkyonides Gulf. It is one of a number of such basins in this area. During the Neogene the basin was under extension and the bounding faults were tectonically active, with abundant evidence of syn-tectonic deformation affecting the contemporaneous sedimentary infill (more than 1 km of alluvial, fluvial, and lacustrine deposits; Bentham et al. 1991). In the Pleistocene (over about the last 1 Ma), the basin bounding faults became inactive and the basin was inverted due to uplift in the footwall of the active fault that bounds the south-east margin of the Alkyonides Gulf. The basin fill was then subjected to weathering, erosion, and the development of drainage networks. Some 9 km3 of footwall sediments have been eroded from the backtilted footwall of the main fault scarp (Figure 1.11). This gives a mean erosion rate of 0.27 mm a−1 and the volume would be sufficient to deposit a layer 180 m thick over the Alkyonides

23

Gulf (some 50 km2 ; Leeder et al. 1991). These rates of erosion are strongly aided by the availability of highly erodible (Plio-Pleistocene basin fill) lithologies. The change in fault activity was probably related to clockwise crustal block rotations of some 25–3◦ that occurred about this time, causing reorientation of the faults from east to west to 115–12◦ (Leeder et al. 1991). This change in tectonic activity was associated with the initiation of new drainage systems and a change in basin slope. Previously the Megara basin general palaeoslope was towards the north-east. Since then it has been back-tilted towards the south-east. This change in slope is recorded in the drainage pattern. The highest order, axial drainages drain to the south-east down the current dip slope (Figure 1.11). The tributaries to these streams are dominantly located to the south-west and flow towards the north-east, suggesting they may be streams originally developed on the northeast dipping palaeoslope (Goldsworthy and Jackson 2000).

Indirect Impact of Tectonics: South East Spain Although many long-term landscape elements of the Mediterranean can be attributed directly to the geodynamics of the region, others are less direct and indicate the complex response of geomorphic systems to external controls. Much of this will be examined in detail in later chapters of this book. Here we focus on a specific process which is abundant throughout the Mediterranean as a function of the regional tectonics—river capture. River capture in its true sense (a bottom-up process) is prolific in areas with high relief, a mixture of dip and strike drainages, and differential uplift. These characteristics provide optimal conditions for river capture to occur i.e. where a lower elevation stream undergoing headward erosion and gullying can breach the drainage divide of a higher elevation drainage system and lead to stream capture. River capture has occurred in response to the Mediterranean drawdown in the Messinian (e.g. Barber 1981) and has even been credited with facilitating the reflooding of the Mediterranean from the Black Sea (Hsü et al. 1973; McCulloch and De Deckker 1989) and Atlantic (Blanc 2002). River capture can lead to sudden, dramatic falls in base level within river systems (for example some 500 m in the Guadix basin, south-east Spain; Calvache and Viseras 1997) with major impacts on sediment flux and routing both within and between drainage systems (e.g. Mather and Harvey 1995; Mather 2000a , b; Mather et al. 2000, 2002; Stokes et al. 2002; Azañón et al. 2005).

24

Anne Mather

River Watershed, Río de Aguas Watershed, Rambla de los Feos R a G mb óc la ha de r

R Cin ambl ta a d Bla e nc a

Rambla de Sorbas/ Río de Aguas

los de bla Ram hopos C

Ra

Badlands

de la mb eos F

Landslides

los

>105 m3 >103 – 105 m3 <103 m3

N

0

5 km

Capture site

Fig. 1.12. Distribution of erosional and landslide features in relation to the 70 ka river capture site (X) in the Río de Aguas basin, south-east Spain. Modified from Griffiths et al. (2002). Note the increase in frequency and magnitude of landslides and greater abundance of badland areas in proximity to the river capture site. The box outlines the area depicted in Figure 1.15.

This has profound impacts on the affected drainage systems, leading to accelerated erosion and landscape instabilities (e.g. badlands and landslides) in the impacted area. The best documented example of the impact of river capture on a fluvial system within the Mediterranean comes from the Aguas/Feos capture (Harvey and Wells 1987; Harvey et al. 1995) of the Sorbas basin of south-east Spain, part of the Betic–Rif system described earlier. Since the late Neogene, compression has dominated in this region with tectonic movement being expressed through differential uplift within and between the sedimentary basins and the mountain ranges. Pliocene to recent average uplift rates are calculated to be in excess of 160 m Ma−1 for the Sierras (Alhamilla/Cabrera), but are typically much lower for the basins (80 m Ma−1 for the Sorbas basin; 11– 21 m Ma−1 for the Vera basin; Mather 2001). This deformation has been significant in generating regional

topographic gradients both within and between basins. The recent high resolution Uranium series dating of pedogenic calcretes by Candy et al. (2005) indicates that the age of the Aguas/Feos capture is c.70 ka. The river capture led to a drop in base level for the Aguas drainage of the Sorbas basin of 90 m (Harvey et al. 1995) and re-routed 73 per cent of the original Sorbas basin drainage (Mather et al. 2002). The modern topography of the Sorbas basin is dominated by an incising drainage network (the Río Aguas) which is associated with landscape instabilities such as extensive, actively eroding badland terrain and landslides (Figure 1.12; Harvey 2001; Mather et al. 2002). Landscape erosion is locally severe, with some areas of abandoned agricultural land undergoing gully headcut retreat of several metres in a single rainstorm event, and rapid development of associated piping. These badland areas are associated with material of low unconfined compressive strength such

Fig. 1.13. Image of an active mass failure along the margins of the Río de Aguas valley near the river capture site discussed in the text. Note the circled people for scale on the ridge crest in the centre of the image. The ridge comprises limestone with large tension cracks on either side indicating the mass failure into the Aguas valley towards the left and right of the image. These failures are facilitated by the rapid incision associated with the 70 ka river capture that occurred at this site (Photo: Anne Mather).

Site A

Site B

(strong lithology)

(weak lithology)

70 ka river capture affects base level Valley width:depth = 3

Late Pleistocene

Valley width:depth = 12

Holocene

Landslides dominate. Rapid valley widening.

Landslide activity still dominant but smaller in magnitude. Badland processes minimal.

Fossil landslides provide template for later badland gully and pipe networks. Badland processes dominate.

Fig. 1.14. Schematic evolution of the valley systems (represented in cross-section) of the Sorbas basin before and after the 70 ka river capture.

26

Anne Mather

as marls and silts. The main landslides, in contrast, tend to dominate in areas of stronger lithologies with higher unconfined compressive strengths such as limestones (Figure 1.13), or in pervious material which in an arid environment has more resistance to erosion, such as gypsum. The landslide features vary in style from deep-seated rotational slips, to block topples with volumes that may exceed 1 million m3 (Hart and Griffiths 1999; Griffiths et al. 2002). Mass movement phenomena from a range of settings are discussed in Chapter 6. The spatial and temporal distribution of the above features in the landscape is governed by their geographic location to the 70 ka river capture site (Figure 1.12). The beheaded river (Rambla de los Feos in Figure 1.12) experienced a reduction in erosion rates. In contrast, the capturing drainage system (Lower Aguas, downstream of the capture site marked on Figure 1.12), incurred dramatic increases in net erosion as a result of the increase in catchment area and stream power. At the capture site the 90 m drop in base level initiated the propagation of a rejuvenating wave of incision up the catchments, accelerating sediment production and delivery to the fluvial system. Over a period of 70 ka this has reached 20 km upstream, at a decaying rate. Near the capture site a >tenfold increase in incision was experienced (Stokes et al. 2002) and this radically altered the sediment delivery processes (Figure 1.14). Initially the generation of steep, rapidly unloaded slopes generated mass movement processes such as landslide

failures. In weaker lithologies the dominant sediment delivery process later became dominated by more progressive slope erosion by surface runoff and subsurface piping processes (pseudokarst). Most of this accelerated erosion is still restricted to the main valley sideslopes, and has not yet reached the main drainage divides so that overall surface lowering is much less than that recorded by the localized valley incision (Figure 1.15). Post-capture channel incision was 50 m about 7 km upstream of the capture site and 25 m about 13 km upstream. In these areas a complex reorganization of tributary drainages through stream capture occurred (Mather 2000a) resulting in extensive badland erosion and the development of landslides of 290,000 m3 to 550,000 m3 in volume (Mather et al. 2003).

Conclusions To understand the individual landscape elements of the Mediterranean basin it is evident that we have to understand the geodynamics of the region on a variety of spatial and temporal scales. Whether examining the long-term evolution of a badland, a karst system, or a river catchment we need to appreciate that many aspects may be inherited from earlier periods of tectonic activity and landscape change such as the Messinian Salinity Crisis. Individual seismic events can lead to catchmentwide landsliding, generating sediment sources for river systems that may take many millennia to be eroded 100

13

80 60

R

12

bl

a

de

40

G

óc

11

Surface lowering (m)

20

ha

r 0 as/ orb e S uas g la d Rb per A Up

10 09 08

R Cin bla d ta Bla e nc a

Modern drainage network

07 06

S

e la d s Rb hopo C los

05 73

74

75

76

77

78

79

80

81

Fig. 1.15. Surface lowering above the 70 ka river capture site depicted in Figure 1.12. Note the highest levels of surface lowering are associated with (1) the valley networks, (2) valley confluences, and overall with (3) the lower reaches of the drainage network proximal to the capture point (which is just off the map to the right). Modified from Mather et al. (2002).

Tectonic Setting and Landscape Development

and transported through the fluvial networks. We also need to appreciate that whilst many processes may be directly attributable to the regional tectonics, such as drainage reversal, the indirect impact of tectonics is not always immediately clear. Indirect impacts, such as river capture, can lead to sudden, high-magnitude changes in geomorphic systems that are still felt in the landscape today, despite (in human terms) their relative antiquity.

References Adams, C. G., Benson, R. H., Kidd, R. B., Ryan, W. B. F., and Wright, R. C. (1977), The Messinian Salinity Crisis and evidence of late Miocene eustatic changes in the world ocean. Nature 269: 383–6. Alexander, D. (1982), Difference between ‘calanchi’ and ‘biancane’ badlands in Italy, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. Geobooks, Norwich, England, 71–88. Alexander, R. W., Harvey, A. M., Calvo, A., James, P. A., and Cerda, A. (1994), Natural stabilisation mechanisms on badland slopes: Tabernas, Almería, Spain, in A. C. Milliman and K. Pye, (eds.), Environmental Change in Drylands: Biogeographical and Geomorphological Perspectives. John Wiley & Sons, Chichester, 85–111. Allen, P. A. (1997), Earth Surface Processes. Blackwell Science, Oxford, 404. Amorosi, A., Farina, M., Severi, P., Preti, D., Caporale, L., and Di Dio, G. (1996), Genetically related alluvial deposits across active fault zones: an example of alluvial fan–terrace correlation from the upper Quaternary of the southern Po Basin, Italy. Sedimentary Geology 102: 275–95. Armijo, R., Meyer, B., King, G. C. P., Rigo, A., and Papanastassiou, D. (1996), Quaternary evolution of the Corinth Rift and its implications for the Late Cenozoic evolution of the Aegean. Geophysical Journal International 126 (1): 11–53. Audra, P., Mocochain, L., Camus, H., Gilli, E., Clauzon, G., and Bigot, J. Y. (2004), The effect of the Messinian Deep Stage on karst development around the Mediterranean Sea. Examples from Southern France. Geodinamica Acta 17: 389–400. Azañón, J. L., Azor, A., Pérez-Peña, J. V., and Carillo, J. M. (2005), Late Quaternary large-scale rotational slides induced by river incision: The Arroyo de Gor area (Guadix basin, SE Spain). Geomorphology 69: 152–68. Barber, P. M. (1981), Messinian subaerial erosion of the protoNile Delta. Marine Geology 44: 253–72. Baumard, B. (2001), Valorisation des données pour l’étude de la crise messinienne dans le Gard rhodanien et la moitié Est del la France. Ph.D. Thesis, École dels Mines de Paris. Benito, G., Gutiérrez, C., and Sancho, C. (1992), Erosion rates in badland areas of the central Ebro Basin (northeast Spain). Catena 19: 269–86. Bentham, P., Collier, R. E. L., Gawthorpe, R. L., Leeder, M. R., Prossor, S., and Stark, C. (1991), Tectono-sedimentary development of an extensional basin: the Neogene Megara basin, Greece. Journal of the Geological Society of London 148: 923–34.

27

Blanc, P.-L. (2002), The opening of the Plio-Quaternary Gibraltar Strait: assessing the size of a cataclysm. Geodinamica Acta 15: 303–17. Briole, P., Rigo, A., Lyon-Caen, H., Ruegg, J. C., Papazissi, K., Mitsakaki, C., Balodimou, A., Veis, G., Hatzfield, D., and Deschamps, A. (2000), Active deformation of the Corinth rift, Greece: Results from repeated Global Positioning System surveys between 1990 and 1995. Journal of Geophysical Research 105: 605–25. Buforn, E., Udias, A., and Madariaga, R. (1991), Intermediate and deep earthquakes in Spain. Pure and Applied Geophysics 136: 375–93. Calaforra, J. M. and Pulido-Bosch, A. (2003), Evolution of the gypsum karst of Sorbas (SE Spain). Geomorphology 50 (1–3): 173–80. Calais, E., DeMets, C., and Nocquet, J. M. (2003), Evidence for a post- 3.16 Ma change in the Nubia-Eurasia plate motion. Earth and Planetary Science Letters 216: 81–92. Calcaterra, D. and Santo, A. (2004), The January 10, 1997 Pozzano landslide, Sorrento Peninsula, Italy. Engineering Geology 75: 181–200. Calvache, M. L. and Viseras, C. (1997), Long-term control mechanisms of stream piracy processes in southeast Spain. Earth Surface Processes and Landforms 22: 93–105. Campbell, I. A. (1989), Badlands and badland gullies, in D.S.G. Thomas (ed.) Arid Zone Geomorphology. Belhaven, London, 158–83. Canals, M., Lastras, G., Urgeles, R., Casamor, J. L., Mienert, J., Cattaneo, A., De Batist, M., Haflidason, H., Imbo, Y., Laberg, J. S., Locat, J., Long, D., Longva, O., Masson, D. G., Sultan, N., Trincardi, F., and Bryn, P. (2004), Slope failure dynamics and impacts from seafloor and shallow sub-seafloor geophysical data: case studies from the COSTA project. Marine Geology 213: 9–72. Candy, I., Black, S., and Sellwood, B. W. (2005), U-Series isochron dating of immature and mature calcretes as a basis for constructing Quaternary landform chronologies for the Sorbas Basin, southeast Spain. Quaternary Research 64: 100–11. Carey, S. W. (1955), The orocline concept in geotectonics. Papers and Proceedings of the Royal Society of Tasmania 89: 255–88. Carminati, E. and Doglioni, C. (2004), Mediterranean Tectonics. Encyclopaedia of Geology, Elsevier, Amsterdam. Carro, M., De Amicis, M., Luzi, L., and Marzrati, S. (2003), The application of predictive modelling techniques to landslides induced by earthquakes: the case study of the 26 September 1997 Umbria-Marche earthquake (Italy). Engineering Geology 69: 139–59. Chester, D. K. and Duncan, A. M. (1982), The interaction of volcanic activity in Quaternary times upon the evolution of the Alcantara and Simento Rivers, Mount Etna, Sicily. Catena 9: 319–42. Clauzon, G. (1982) The Messinian Rhône canyon as a definite proof of the “desiccated deep-basin model”. Bulletin Société Géologique de France 24: 597–610. Suc, J.-P., Gautier, F., Berger, A., and Loutre, M.-F. (1996), Alternate interpretation of the Messinian Salinity Crisis: Controversy resolved? Geology 24: 363–6. Collier, R. E. L., Leeder, M. R., and Jackson, J. A. (1995), Quaternary drainage development, sediment fluxes and extensional tectonics in Greece, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 31–44.

28

Anne Mather

Comas, M. C., Platt, J. P., Soto, J. I., and Watts, A. B. (1999), The origin and tectonic history of the Alboran Basin. Proceedings of the ODP Science Research 161: 555–80. Corominas, J. and Moya, J. (1999), Reconstructing recent landslide activity in relation to rainfall in the Llobregat River basin, Eastern Pyrenees, Spain. Geomorphology 30: 79–93. Darawcheh, R., Sheinati, M. R., Margottini, C., and Paolini, S. (2000), The 9 July 551 AD Beirut earthquake, Eastern Mediterranean Region. Journal of Earthquake Engineering 4: 403–14. De Ploey, J. (1974), Mechanical properties of hillslopes and their relation to gullying in central semi-arid Tunisia. Zeitschrift für Geomorphologie 21: 177–90. Dewey, J. F. (1980) Episodicity, sequence and style at convergent plate boundaries. Geological Association of Canada Special Paper 20: 553–73. Doglioni, C., Innocenti, F., and Mariotti, G. (2001), Why Mt Etna? Terra Nova 13: 25–31. Droz, L. (1983), L’Eventail sous-marin profound du Rhône (Golfed u Lion): grands traits morphologiques et structure semi-profonde. 3rd cycle thesis, University of Paris VI. Duermeijer, C. E., Nyst, M., Meijer, P. T., Langereis, C. G., and Spakman, W. (2000), Neogene evolution of the Aegean arc: palaeomagnetic and geodetic evidence for a rapid and young rotation phase. Earth and Planetary Science Letters 176: 509–25. Duggen, S., Hoernle, K., Van den Bogaard, P., Rüpke, L., and Morgan, J. P. (2003), Deep roots of the Messinian Salinity Crisis. Nature 422: 602–6. Durand Delga, M. (1980), La méditerranée occidentale étapes de sa genèse et problèmes structuraux liess a celle-ci. Memoire, Société Géologique de France 10: 203–24. El-Robrini, M., Gennesseaux, M., and Mauffret, A. (1985), Consequences of the El-Asnam earthquakes: turbidity currents and slumps on the Algerian Margin (Western Mediterranean). Geo. Marine Letters 5: 171–6. Elsasser, W. M. (1971), Sea floor spreading as thermal convection. Journal of Geophysical Research 76: 1101–12. Faccenna, C., Mattei, M., Funiciello, R., and Jolivet, L. (1997), Styles of backarc extension in the central Mediterranean. Terra Nova 9: 126–30. Ferrarese, F., Macaluso, T., Madonia, G., Palmeri, A., and Sauro, U. (2003), Solution and recrystallisation processes and associated landforms in gypsum outcrops of Sicily. Geomorphology 49 (1–2): 25–43. Fiorillo, F. and Wilson, R. C. (2004), Rainfall induced debris flows in pyroclastic deposits, Campania (southern Italy). Engineering Geology 75: 263–89. Fortuin, A. R., Krijgsman, W., Hilgen, F. J., and Sierro, F. J. (2000), Late Miocene Mediterranean desiccation: topography and significance of the ‘Salinity Crisis’ erosion surface onland on southeast Spain: Comment. Sedimentary Geology 133: 167–74. Galadini, F. and Galli, P. (2003), Palaeoseismology of silent faults in the central Apennines (Italy): the Mt. Vettore and Laga Mts. Faults. Annals of Geophysics 46: 815–36. Gallart, F., Llorens, P., Latron, J., and Regues, D. (2002), Hydrological processes and their seasonal controls in a small Mediterranean mountain catchment in the Pyrenees. Hydrology and Earth System Sciences 6: 527–37. Garcés, M., Krigsman, W., and Agusti, J. (1998), Chronology of the late Turolian deposits of the Fortuna Basin (SE Spain):

implications for the Messinian evolution of the eastern Betics. Earth and Planetary Science Letters 163: 69–81. (2001), Chronostratigrapic framework and evolution of the Fortuna basin (Eastern Betics) since the late Miocene. Basin Research 13: 199–216. Gargani, J. (2004), Modelling of the erosion in the Rhône valley during the Messinian crisis (France). Quaternary International 121: 13–22. Gaullier, V., Anronnini, E., Benkhelil, J., and Got, H. (1998), Corps gravitaires récents dans le basin Nord-Baléares: géométrie et quantification. C.R. Acad. Sci. Paris, Sér. IIA 327: 677–84. Gautier, P., Brun, J. P., Moriceau, R., Sokoutis, D., Martinod, J., and Jolivet, L. (1999), Timing, kinematics and cause of Aegean extension: a scenario based on a comparison with simple analogue experiments. Tectonophysics 315: 31–72. Goes, S., Giardini, D., Jenny, S., Hollenstein, C., Kahle, H. G., and Geiger, A. (2004), A recent tectonic reorganization in the south-central Mediterranean. Earth and Planetary Science Letters 226: 335–45. Goldsworthy, M. and Jackson, J. (2000), Active normal fault evolution in Greece revealed by geomorphology and drainage patterns. Journal of the Geological Society of London 157: 967–81. Griffiths, J. S., Mather, A. E., and Hart, A. B. (2002), Landslide susceptibility in the Rio Aguas catchment, SE Spain. Quarterly Journal of Engineering Geology and Hydrogeology 35: 9–17. Grove, A. T. and Rackham, D. (2001), The Nature of Mediterranean Europe: An Ecological History. Yale University Press, New Haven, Conn. Gueguen, E., Doglioni, C., and Ferdinández, M. (1997), Lithospheric boudinage in the western Mediterranean back-arc basins. Terra Nova 9 (4): 184–7. (1998), On the post-25 Ma geodynamic evolution of the western Mediterranean. Tectonophysics 298: 259–69. Gutscher, M. A., Malod, J., Rehault, J. P., Contrucci, I., Klingelhoefer, F., Mended-Victor, L., and Spakman, W. (2002), Evidence for active subduction beneath Gibraltar. Geology 30: 1071–4. Harris, D. and Vita-Finzi, C. (1968) Kokkinopilos—a Greek badland. Geographical Journal 134: 537–46. Hart, A. B. and Griffiths, J. S. (1999), Mass movement features in the vicinity of the town of Sorbas, Southeast Spain, in J. S. Griffiths, R. B. Thomas, and M. Stokes (eds.), Landslides. Proceedings of the 10th International Field Conference on Landslides, Balkema, Rotterdam, 57–63. Harvey, A. (1982), The role of piping in the development of badlands and gully systems in south-east Spain, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. Geobooks, Norwich, 317–35. (2001), Uplift, dissection and landform evolution: the Quaternary, in A. E. Mather, J. M. Martin, A. M. Harvey, and J. C. Braga (eds.), A Fieldguide to the Neogene Sedimentary Basins of the Almería Province, SE Spain. Blackwell Science, Oxford, 225–322. Harvey, A. and Wells, S. G. (1987), Response of Quaternary fluvial systems to differential epeirogenic uplift: Aguas and Feos River systems, south-east Spain. Geology 15: 689–93. Miller, S. Y. and Wells, S. G. (1995), Quaternary soil and river terrace sequences in the Aguas/Feos river systems: Sorbas basin, southeast Spain, in J. Lewin, M. G. Macklin, and

Tectonic Setting and Landscape Development J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 263–81. Goy, J., Mather, A. E., Silva, P., Stokes, M., and Zazo, C. (1999), The impact of Quaternary sea level and climate change on coastal alluvial fans in the Cabo de Gata ranges, southeast Spain. Geomorphology, 28: 1–22. Mather, A. E. and Stokes, M. (2005), Alluvial fans: geomorphology, sedimentology, dynamics—introduction. A review of alluvial fan research, in A. M. Harvey, A. E. Mather, and M. Stokes (eds.), Alluvial Fans: Geomorphology, Sedimentology, Dynamics. Geological Society of London Special Publication 251: 1–7. Heezen, B. C. and Ewing, M. (1955), Orleansville earthquake and turbidity currents. AAPG Bulletin 39: 2505–14. Hodell, D. A., Elmstron, K. M., and Kennet, J. P. (1986), Latest Miocene benthic ‰18 O changes, global ice volume, sea level and the ‘Messinian Salinity Crisis’. Nature, 320: 411–14. Benson, R. H., Kennet, J. P., and El Bied, K. R. (1989), Stable isotope stratigraphy of Latest Miocene sequences in northwest Morocco: The Bou Regreg section. Palaeogeography 4: 467–82. Curtis, J. H., Sierro, F. J., and Raymo, M. E. (2001), Correlation of late Miocene to early Pliocene sequences between the Mediterranean and North Atlantic. Palaeogeography 16: 164–78. Houseman, G. (1996), From mountains to basin. Nature 379: 771–2. Hsü, K. J. and Bernoulli, D. (1978), Genesis of the Tethys and the Mediterranean. Initial Report DSDP 42: 943–50. Ryan, W. B. F., and Cita, M. B. (1973), Late Miocene desiccation of the Mediterranean, Nature 242: 240–4. Montadert, L., Bernoulli, D., Cita, M. B., Erickson, A., Garison, R. E., Kidd, R. E., Melieres, R. B., Muller, C., and Wright, R. (1977), History of the Messinian Salinity Crisis. Nature 267: 399–403. Imeson, A. C., Kwaad, F. J. P. M., and Verstraten, J. M. (1982), The relationship of soil physical and chemical properties to the development of badlands in Morocco, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. Geobooks, Norwich, 47–70. Jackson, J. A., Gagnepain, J., Houseman, G., King, G. C. P., Papadimitriou, P., Soufleris, C., and Virieux, J. (1982), Seismicity, normal faulting, and the geomorphological development of the Gulf of Corinth (Greece): The Corinth earthquake of February and March 1981. Earth and Planetary Science Letters 57: 377–97. Jolivet, I. and Faccenna, C. (2000), Mediterranean extension and the Africa-Eurasian collision. Tectonics 19: 1095–106. Josnin, J. Y. (2001), Quaternary tectonics: influence on the structure of two karstic aquifers of Languedoc, France. Eclogae Geologicae Helvetiae 94: 173–84. Kahle, H. G., Straub, C., Reilinger, R., McClusky, S., King, R., Hurst, K., Vries, G., Kastens, K., and Cross, P. (1998), The strain rate field in the eastern Mediterranean region, estimated by repeated GPS measurements. Tectonophysics 294: 237–52. Keefer, D. K. (1994), The importance of earthquake-induced landslides to long-term slope erosion and slope-failure hazards in seismically active regions. Geomorphology 10: 265–84. (2002), Investigating landslides caused by earthquakes—a historical review. Surveys in Geophysics 23: 473–510.

29

Keraudren, B. and Sorel, D. (1987), The terraces of Corinth (Greece)—a detailed record of eustatic sea-level variations during the last 500,000 years. Marine Geology 77: 99–107. King, G., Sturdy, D., and Bailey, G. (1997), The tectonic background to the Epirus landscape, in G. N. Bailey (ed.), Klithi: Palaeolithic settlement and Quaternary landscapes in northwest Greece, ii. Klithi in its local and regional setting. McDonald Institute Monographs, Cambridge, 541–58. Krijgsman, W. and Garcés, M. (2004), Palaeomagnetic constraints on the geodynamic evolution of the Gibraltar Arc. Terra Nova 16: 281–7. Langereis, C. G., Zacharariasse, W. J., Bocaletti, M., Moriatti, G., Gelati, R., Iaccarino, S., Papani, G., and Villa, G. (1999a ), Late Neogene evolution of the Taza-Guercif Basin (Rifian Corridor, Morocco) and implications for the Messinian Salinity Crisis. Marine Geology 153: 147–60. Hilgen, F. J., Raffi, I., Sierro, F. J., and Wilson, D. S. (1999b), Chronology, causes and progression of the Messinian Salinity Crisis. Nature 400: 652–5. Garcés, M., Agusti, J., Raffi, I., Taberner, C., and Zachariasse, W. J. (2000), The ‘Tortonian salinity crisis’ of the eastern Betics (Spain). Earth and Planetary Science Letters 181: 497–511. Lastras, G., Canals, M., Urgeles, R., De Bastist, M., Calafat, A. M., and Casamor, J. L. (2004), Characterisation of the recent BIG’95 debris flow deposit on the Ebro margin, western Mediterranean Sea, after a variety of seismic reflection data. Marine Geology 213: 235–55. Leeder, M. R., Seger, M. J., and Stark, C. P. (1991), Sedimentation and tectonic geomorphology adjacent to major active and inactive normal faults, southern Greece. Journal of the Geological Society of London 148: 331–43. Collier, R. E. L., Abdul Aziz, L. H., Trout, M., and Ferentinos, G. (2002), Tectono-sedimentary processes along an active marine/lacustrine half-graben margin: Alkyonides Gulf, Eastern Gulf of Corinth, Greece. Basin Research 14: 25–41. McNeill, L. C., Collier, R. E. L., Portman, C., Rowe, P. J., Andrews, J. E., and Gawthorpe, R. L. (2003), Corinth rift margin uplift: New evidence from Late Quaternary marine shorelines. Geophysical Research Letters 30 (12): Article 1. Lenci, F., Carminati, E., Doglioni, C., and Scrocca, D. (2004), Basal décollement and subduction depth versus topography in the Apennines–Calabrian arc. Boll. Soc. Geol. It. 123: 497–502. Le Pichon, X. and Angelier, J. (1979), The Hellenic arc and trench system: a key to the neotectonic evolution of the eastern Mediterranean area. Tectonophysics 60: 1–42. Pautot, G., Auzende, J. M., and Olivet, J. L. (1971), La Méditerranée occidentale. Biochronologie, correlations avec les formations marines et echanges intercontinenteaux. Earth and Planetary Science Letters 13: 145–52. Lister, G. S., Forster, M. A., and Rawling, T. (2001), Episodicity during orogenesis, in J. A. Miller, R. E. Holdworth, I. S. Buick, and M. Hand (eds.), Continental Reactivation and Reworking. Geological Society of London Special Publication, 89–113. Lonergon, L. and White, N. (1997), Origin of the Betic-Rif mountain belt. Tectonics 16: 504–22. Lucente, P. F., Chirarabba, C., Cimini, G. B., and Giardini, D. (1999), Topographic constraints on the geodynamic evolution

30

Anne Mather

of the Italian region. Journal of Geophysical Research 104: 20307–27. Luino, F. (2005), Sequence of instability processes triggered by heavy rainfall in northern Italy. Geomorphology 66: 13–39. McClusky S., Balassanian, S., Barka, A., Demir, C., Ergintav, S., Georgiev, I., Gurkan, O., Hamburger, M., Hurst, K., Khale, H., Kastens, K., Kekelidze, G., King, R., Kotzev, V., Lenk, O., Mahmoud, S., Mishin, A., Nadariya, M., Ouzounis, A., Paradissis, D., Peter, Y., Prilepin, M., Reilinger, R., Sanli, I., Seeger, H., Tealeb, A., Toksšz, M. N., and Veis G. (2000), Global positioning system constraints on plate kinematics and dynamics in the eastern Mediterranean and Caucasus. Journal of Geophysical Research 105: 5695–719. McCulloch, M. T. and De Deckker, P. (1989), Sr isotope constraints on the Mediterranean environment at the end of the Messinian Salinity Crisis. Nature 342: 62–5. Macklin, M. G., Lewin, J., and Woodward, J. C. (1995), Quaternary fluvial systems in the Mediterranean Basin, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 1–25. Malinverno, A. and Ryan, W. B. F. (1986), Extension of the Tyrrhenian Sea and shortening in the Apennines as a result of arc migration driven by sinking of the lithosphere. Tectonics 5: 227–45. Cafiero, M., Ryan, W. B. F., and Cita, M. B. (1981), Distribution of the Messinian sediments and erosional surfaces beneath the Tyrrhenian Sea: geodynamic implications. Oceanologica Acta 4: 489–95. Maramai, A., Graziani, L., and Tinti, S. (2005a ), Tsunamis in the Aeolian islands (southern Italy): a review. Marine Geology 215: 11–21. Alessro, G., Burrato, P., Colini, L., Cucci, L., Nappi, R., Nardi, A., and Vilardo, G. (2005b), Near and far-field survey report of the 30th December 2002 Stromboli (Southern Italy) tsunami. Marine Geology 215: 93–106. Marshak, S. (1988), Kinematics of orocline and arc formation in thin-skinned orogens. Tectonics 7: 73–86. Mather, A. E. (2000a ), Adjustment of a drainage network to capture induced base-level change. Geomorphology 34: 271–89. (2000b), Impact of headwater river capture on alluvial system development. Journal of the Geological Society of London 157: 957–66. (2001), The development of the Neogene Basins, in A. E. Mather, J. M. Martin, A. M. Harvey, and J. C. Braga (eds.), A Fieldguide to the Neogene Sedimentary Basins of the Almería Province, SE Spain. Blackwell Science, Oxford, 29–133. and Harvey, A. M. (1995), Controls on drainage evolution in the Sorbas Basin, SE Spain, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 65–76. and Stokes, M. (2000), Quantifying long-term catchment changes of alluvial fan systems. Geological Society of America Bulletin 112: 1825–33. Stokes, M., and Griffiths, J. S. (2002), Quaternary landscape evolution: a framework for understanding contemporary erosion, SE Spain. Land Degradation and Management 13: 1–21. Griffiths J. S., and Stokes M. (2003), Anatomy of a ‘fossil’ landslide from the Pleistocene of SE Spain. Geomorphology 50: 135–49.

Martinez-Diaz, J. J., Masana, E., Hernandez-Enrile, J. L., and Santanach, P. (2003), Effects of repeated palaeoearthquakes on the Alhama de Murcia Fault (Betic Cordillera, Spain) on the Quaternary evolution of an alluvial fan system. Annals of Geophysics 46: 775–91. Mauffret, A. (1979), Etude geodynamique de la merge des iles Baleares. Memoirs Société geologie France, N.S. LVI (132), 94. Méar, Y. (1984), Séquences et unites sédimentaires du glacis rhodanien (Méditerranée Occidentale). 3rd cycle thesis, Perpignan University, France. 214 pp. Michetti, A. M., Ferreli, L., Serva, L., and Vittori, E. (1997), Geological evidence for strong historical earthquakes in an ‘aseismic’ region: the Pollino case (Southern Italy). Journal of Geodynamics 24: 67–86. Miller, K. G., Wright, J. D., and Fairbanks, R. G. (1991), Unlocking the ice house: Oligocene-Miocene isotopes, eustasy and margin erosion. Journal of Geophysical Research 96: 6829–48. Molnar, P. and Atwater, T. (1978), Interarc spreading and the Cordilleran tectonics as alternates related to the age of the subducted oceanic lithosphere. Earth and Planetary Science Letters 41: 330–40. Murphy, W. (1995), The geomorphological controls on seismically triggered landslides during the 1908 Straits of Messinia Earthquake, Southern Italy. Quarterly Journal of Engineering Geology 28: 61–74. Nicoletti, P. G. and Parise, M. (2002), Seven landslide dams of old seismic origin in southeastern Sicily (Italy). Geomorphology 46: 203–22. Nocquet, J. M. and Calais, E. (2004), Geodetic measurements of crustal deformation in the western Mediterranean and Europe. Pure and Applied Geophysics 161: 661–81. Norman, S. E. and Chase, C. G. (1983), Uplift of the shores of the western Mediterranean due to Messinian desiccation and flexural isostasy. Nature 322: 450–1. Ollier, C. D. (1995), Tectonics and landscape evolution in southeast Australia. Geomorphology 12: 37–44. Orri, G. G. (1993), Continental depositional systems of the Quaternary of the Pop Plain (Northern Italy). Sedimentary Geology 83: 1–14. Ozer, A. and Vita-Finzi, C. (eds.) (1986), Dating Mediterranean Shorelines. Zeitschrift für Geomorphologie, Suppl. 62. Papadopoulos, G. A. and Plessa, A. (2000), Magnitude–distance relations for earthquake-induced landslides in Greece. Engineering Geology 58: 377–86. Karastathis, V. K., Ganas, A., Pavlides, S., Fokaefs, A., and Orfanogiannaki, K. (2003), The Lefkada, Ionian Sea (Greece), shock (Mw 6.2) of 14 August 2003: Evidence for the characteristic earthquake from seismicity and ground failures. Earth, Planets and Space 55: 713–18. Pino, N. A., Ripepe, M., and Cimini, G. B. (2004), The Stromboli Volcano landslides of December 2002: a seismological description. Geophysical Research Letters 31 L02605, doi:10.1029/2003GL018385. Platt, J. P. and Vissers, R. L. M. (1989), Extensional collapse of thickened continental lithosphere: a working hypothesis for the Alboran Sea and Gibraltar arc. Geology 17: 540–3. Porfido, S., Esposito, E., Vittori, E., Tranfaglia, G., Michetti, A. M., Blumetti, M., Fereli, L., Guerrieri, L., and Serva, L. (2002),

Tectonic Setting and Landscape Development Aerial distribution of ground effects induced by strong earthquakes in the Southern Apennines (Italy). Surveys in Geophysics 23: 529–62. Prinz, D., Gomer, D., and Belz, S. (1994), Studies of the causes of soil-erosion on marl soils in Northern Algeri—the role of traditional soil tillage. Land Degradation and Rehabilitation 5: 271–80. Rehault, J.-P., Boillet, G., and Mauffret, A. (1985), The Western Mediterranean Basin, in D. J. Stanley and F. C. Wezel (eds.), Geological Evolution of the Mediterranean Basin. Springer, New York, 101–29. Reicherter, K. R., Jabaloy, A., Galindo-Zaldivar, J., Ruano, P., Becker-Heidmann, P., Morales, J., Reiss, S., and Gonzalez-Lodeiro, F. (2003), Repeated palaeoseismic activity of the Ventas de Zafarraya fault (S. Spain) and its relation with the 1884 Andalusian earthquake. International Journal of Earth Sciences 92: 912–22. Riding, R., Braga, J. C., and Martin, J. M. (2000), Late Miocene Mediterranean desiccation: topography and significance of the ‘Salinity Crisis’ erosion surface on-land on southeast Spain: Reply. Sedimentary Geology 133: 175–84. Rodriguez, C. E., Bommer, J. J., and Chandler, R. J. (1999), Earthquake-induced landslides: 1980–1997. Soil Dynamics and Earthquake Engineering 18: 325–46. Rosenbaum, G. and Lister, G. S. (2002), Reconstruction of the evolution of the Alpine–Himalayan orogen—an introduction, in G. Rosenbaum and G. S. Lister (eds.), Reconstruction of the Evolution of the Alpine-Himalayan Orogen. Journal of the Virtual Explorer 8: 1–2. (2004a ), Neogene and Quaternary rollback evolution of the Tyrrhenian Sea, the Apennines and the Sicilian Maghrebides. Tectonics 23. (2004b), Formation of arcuate orogenic belts in the western Mediterranean region. Geological Society of America Special Paper 383: 41–56. and Duboz, C. (2002a ), Reconstruction of the tectonic evolution of the western Mediterranean since the Oligocene. Journal of the Virtual Explorer 8: 107–30. (2002b), Relative motions of Africa, Iberia and Europe during the Alpine Orogeny. Tectonophysics 359: 117–29. Rothwell, R. G., Thomson, J., and Kähler, G. (1998), Low sea-level emplacement of a very large Late Pleistocene ‘megaturbidite’ in the western Mediterranean Sea. Nature 392: 377–80. Reeder, M. S., Anastasakis, G., Stow, D. A. V., Thomson, J., and Kähler, G. (2000), Low-stand emplacement of megaturbidies in the Western and Eastern Mediterranean Sea. Sedimentary Geology 135: 75–88. Royden, L. H. (1993a ), Evolution of retreating subduction boundaries formed during continental collision. Tectonics 12: 629–38. (1993b), The tectonic expression of slab pull at continental convergent boundaries. Tectonics 12: 303–25. Ryan, W. B. F. (1978), Messinian badlands on the southeastern margin of the Mediterranean Sea. Marine Geology 27: 349–63. and Cita, M. B. (1978), The nature and distribution of Messinian erosional surfaces—indicators of a severalkilometre-deep Mediterranean in the Miocene. Marine Geology 27: 193–230.

31

Salustri Galli, C., Torrini, A., Doglioni, C., and Scrocca, D. (2002), Divide and highest mountains versus subduction in the Apennines. Studi Geologici Camerti 1: 143–53. Schlupp, A., Caluzon, G., and Avouac, J.-P. (2001), PostMessinian movement along the Nimes Fault: implication for the seismotectonics of Provence (France). Bulletin Societe Géologique de France 172: 697–711. Silva, P. G., Harvey, A. M., Zazo, C., and Goy, J. L. (1992), Geomorphology, depositional style and morphometrics of alluvial fans in the Guaalentin depression (Murcia, Southeast Spain). Zeitschrift für Geomorphologie 36: 325–41. Goy, J. L., Zazo, C., and Bardaji, T. (2003), Fault-generated mountain fronts in southeast Spain: geomorphologic assessment of tectonic and seismic activity. Geomorphology 50: 203–25. Audemard, F. A., and Mather, A. E. (eds.) (2008), Impact of active tectonics and uplift on fluvial landscapes and drainage development. Geomorphology 102: 1–204. Soria, J. M., Fernández, J., and Viseras, C. (1999), Late Miocene stratigraphy and palaeogeographic evolution of the intramontane Guadix Basin (Betic Cordillera): implications for an Atlantic-Mediterranean connection. Palaeogeography, Palaeoclimatology, and Palaeoecology 151: 255–66. Soudet, H. J., Sorriaux, P., Rolando, J. P. (1994), Relationship between fractures and karstification—the oil bearing palaeokarst of Rospo Mare, Italy. Bulletin des Centres de Recherches Exploration-Production Elf Aquitaine 18: 257–97. Stokes, M., Mather, A. E., and Harvey, A. M. (2002), Quantification of river capture induced base-level changes and landscape development, Sorbas Basin, SE Spain, in S. J. Jones and L. E. Frostick (eds.), Sediment Flux to Basins: Causes, Controls and Consequences. Geological Society of London Special Publication 191: 23–35. Thornes, J. B. and Alcantara-Ayala, I. (1998), Modelling mass failure in a Mediterranean mountain environment: climatic, geological, topographical and erosional controls. Geomorphology 24: 87–100. Vannucci, G., Pondrelli, S., Argnani, A., Morelli, A., Gasperini, P., and Boschi, E. (2004), An Atlas of Mediterranean seismicity. Annals of Geophysics 47: 247–306. Vita-Finzi, C. and King, G. C. (1985), The seismicity, geomorphology and structural evolution of the Corinth area of Greece. Philosophical Transactions of the Royal Society of London A314: 379–407. Warny, A. A., Bart, P. J., and Suc, J.-P. (2003), Timing and progression of the climatic, tectonic and glacioeustatic influences on the Messinian Salinity Crisis. Palaeogeography, Palaeoclimatology, Palaeoecology 202: 59–66. Wasowski, J. and Del Gaudio, V. (2000), Evaluating seismically induced mass movement hazard in Caramanico Terme (Italy). Engineering Geology 58: 291–311. Weijermars, R. (1988), Neogene tectonics in the Western Mediterranean may have caused the Messinian Salinity Crisis and an associated glacial event. Tectonophysics 148: 211–19. Woodward, J. C. (1995a ), Archaeology and human-river environment interactions, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema Rotterdam, 99–102. (1995b), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster,

32

Anne Mather

A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water Quality in River Catchments. John Wiley, Chichester, 365–89. Lewin, J., and Macklin, M. G. (1992), Alluvial sediment sources in a glaciated catchment: the Voidomatis Basin, northwest Greece. Earth Surface Processes and Landforms 17: 205–16.

Wortel, M. J. R. and Spakman, W. (2000), Subduction and slab detachment in the Mediterranean–Carpathian region, Science 290: 1910–17. Yair, A., Goldberg, P., and Brimer, B. (1982), Long term denudation rates in the Zin-Havarim badlands, northern Negev, Israel, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. Geobooks, Norwich, 279–91.

This chapter should be cited as follows Mather, A. E. (2009), Tectonic setting and landscape development, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 5–32.

2

The Marine Environment: Present and Past Eelco Rohling, Ramadan Abu-Zied, James Casford, Angela Hayes, and Babette Hoogakker

Introduction The Mediterranean is a landlocked, semi-enclosed marginal sea that spans a maximum of 3,860 km in the west–east direction, and a maximum of ∼1, 600 km in the north–south direction. Along its roughly 46,000 km of coastline, the basin is enclosed by mountainous terrain, except for a part of the North African margin to the east of Tunisia. The Mediterranean Sea contains very deep basins, more than 4 km, and has an average depth of approximately 1,500 m. Its only natural connection with the open (Atlantic) ocean is through the narrow Strait of Gibraltar, which contains a 284-m deep sill (at a width of ∼30 km), and reaches a minimum width of only 14 km (at a depth of 880 m) (Bryden and Kinder 1991). The Strait of Sicily subdivides the Mediterranean Sea into a western and an eastern basin. This strait is relatively wide (about 130 km) and contains a topographically complex sill-structure with an estimated average depth of 330 m (Wüst 1961), reaching 365 and 430 m in the two major channels (Garzoli and Maillard 1979). The eastern Mediterranean contains two smaller marginal basins, namely the Adriatic Sea and the Aegean Sea (Figure 2.1). Watermasses are exchanged through both the Strait of Gibraltar and the Strait of Sicily by eastward surface and westward subsurface flows (Figure 2.2). This pattern of exchange results from a net buoyancy loss in the basins on the easterly side of the sills, primarily due to strong net evaporative loss from the Mediterranean, and secondarily to some net cooling. Deep

water ventilation in the Mediterranean is primarily salt-driven, and secondarily temperature-driven. This is similar to the mode observed in the present-day Red Sea, but contrasts with the temperature-dominated mode in the modern world ocean. As such, the Mediterranean deep ventilation might be more appropriately described as halo-thermal rather than with the common term thermo-haline. This offers a useful analogue for world ocean circulation modes in past times with very warm and relatively equable global climates, such as the Mesozoic. Interestingly, the Mediterranean is characterized by periodic, widespread deposition of organic-rich sediments or ‘sapropels’ over periods of several thousands of years, similar (in miniature) to the deposition of ‘black shales’ in the Mesozoic oceans. Surface water flowing in through the Strait of Gibraltar is traceable through the Strait of Sicily into the eastern Mediterranean, although its salinity increases steadily towards the east (e.g. Wüst 1961; MalanotteRizzoli and Hecht 1988; Malanotte-Rizzoli and Bergamasco 1989; Pinardi and Masetti 2000) (Figures 2.2 and 2.3). The eastward salinity increase culminates in values around 39.2 psu (up to an extreme of 39.5 psu, Wüst 1960) in the eastern Levantine sector of the Mediterranean, compared with 36.1–36.2 psu for the Atlantic inflow at Gibraltar. The high Levantine salinities are associated with high temperatures in summer, but strong winter cooling (especially between Cyprus and Rhodes) causes surface waters to attain high enough densities to sink and spread at intermediate depths (150–600 m). This forms the

34

Eelco Rohling et al.

Fig. 2.1. Map of the Mediterranean Sea.

Fig. 2.2. Longitudinal cross-section showing water mass circulation in the Mediterranean Sea during the present-day winter (modified from Wüst 1961). Isolines indicate salinity values in psu (practical salinity units) and arrows indicate the direction of water circulation in the Mediterranean Sea.

‘Levantine Intermediate Water (LIW)’. This watermass spreads westward from its formation area throughout the entire Mediterranean Sea. Admixtures of regional winter mixed-layer waters slightly reduce the LIW salinity as it spreads, transforming this watermass into what has become known as ‘Mediterranean Intermediate

Water (MIW)’. There are also contributions of Eastern Mediterranean Deep Water (EMDW) and Western Mediterranean Deep Water (WMDW) to the MIW upon its passage through the Strait of Sicily and the Strait of Gibraltar. In most parts of the eastern Mediterranean, MIW salinities are between 38.8 and 39.1 psu, while

The Marine Environment: Present and Past

35

Fig. 2.3. Surface water circulation in the Mediterranean Sea (modified from Vergnaud-Grazzini et al. 1988; Roussenov et al. 1995). Shaded areas indicate intermediate and deep water formation.

values in the western Mediterranean are between 38.5 and 38.8 psu. The subsurface outflow from the Mediterranean through the Strait of Gibraltar has a salinity of 38.2–38.4 psu (among many others: Wüst 1960, 1961; Garzoli and Maillard 1979; Gascard and Richez 1985; Bryden et al. 1994). The influence of Mediterranean Outflow can be traced as a salinity maximum centred on about 1,000 m depth in the North Atlantic (e.g. Reid 1979; Hill and Mitchelson-Jacob 1993; Iorga and Lozier 1999, O’NeillBaringer and Price 1999). This maximum represents the overall average signature, but an important component of the dispersal of Mediterranean Outflow within the North Atlantic has been found to occur in the form of discrete subsurface ‘lenses’ of salty and warm Mediterranean water. These are the so-called Mediterranean eddies or ‘Meddies’ with diameters up to 100 km, the pathways of which have been traced with neutralbuoyancy floats (Richardson et al. 1991, 2000). The isopycnals (lines of equal water density) at which Mediterranean outflow settles show northward shoaling within the north-east Atlantic. Near the Iceland– Scotland Ridge deep winter mixing of fresher surface waters with the salty Mediterranean tongue raises the salinity of the surface water that enters the Norwegian Sea through the Faroe–Shetland Channel (Hill and

Mitchelson-Jacob 1993). This preconditions the inflow for later convection by increasing its salinity by several tenths of a psu relative to ‘background’ (Reid 1979), the density equivalent of 1–2◦ C cooling. Such preconditioning may facilitate the formation of North Atlantic Deep Water (NADW) in the Norwegian Sea (Reid 1979; Hill and Mitchelson-Jacob 1993). Returning attention to the Mediterranean now, EMDW and WMDW are found below about 1 km depth in the eastern and western Mediterranean basins, respectively, separated by the sill in the Strait of Sicily. Between about 600 and about 1,000 m, a transitional watermass is found between the deep waters and MIW. WMDW is formed in the northern sector of the western Mediterranean, notably in the Gulf of Lions, due to strong winter cooling caused by cold continental air outbreaks that are orographically channelled towards the basin via the Rhône valley (the ‘Mistral’). EMDW is formed in two separate regions, namely the Adriatic Sea and the Aegean Sea. Both areas are subject to orographically channelled continental air outbursts in winter, the ‘Bora’ over the Adriatic, and the ‘Vardar’ over the Aegean Sea (Chapter 3). In schematic terms, the Mediterranean deep water ventilation can be viewed as a two-stage motor (a detailed explanation is given below). The first stage

36

Eelco Rohling et al.

consists of the dominantly salt-driven formation of LIW. The salt distributed throughout the Mediterranean Sea by LIW/MIW preconditions the basin for deep water formation. The second stage of the deep-ventilation motor is dominated by cooling events related to orographically channelled continental air outbursts over northern sectors of the basin in winter. Given the presence of a major monsoon-fed river (Nile) in close proximity to the centre of action of the salt-driven first stage of the motor, we can expect that monsoon variations would be reflected in the efficiency of the Mediterranean’s deep ventilation, as deep water preconditioning would be directly affected. However, we can also expect important northerly climate impacts on the deep ventilation, related to changes in the frequency and/or intensity of winter cooling events. The small volume of the Mediterranean Sea, compared with ocean basins, causes changes in its climatic forcing to be recorded virtually instantaneously in palaeoceanographic proxy data, such as stable isotope and other geochemical ratios, and microfossil abundances. The basin’s limited communication with the open ocean implies that any climatic signals will be recorded in an amplified fashion in Mediterranean properties, such as salinity and specific elemental concentrations. The critical location of the Mediterranean Sea on the boundary between a subtropical/monsoon regime and the temperate westerlies means that it is highly sensitive to changes in both these systems. Since both systems primarily affect fundamentally different characteristics of the basin, the Mediterranean is an excellent site for study of the relative timing and impact of changes in the two major systems (subtropical/monsoon climate predominantly affects freshwater balance, while the temperate westerly climate controls cooling in the north). The regularly recurring deposition of organic-rich ‘anoxic’ sapropels offers discrete windows for very high-resolution study, since these intervals are not affected by sediment homogenization due to bioturbation. The Mediterranean Sea may hold important clues as to the functioning of circulation in the Mesozoic oceans and the formation processes of black shales that are of great economic importance. To appreciate fully the processes underlying past changes in Mediterranean climate and hydrography, comprehensive background knowledge of present-day conditions is indispensable. Following a brief history of the development of the Mediterranean basin, therefore, this chapter first discusses relevant aspects of the region’s modern climate and oceanography before

engaging in a review of palaeoclimatological and palaeoceanographic reconstructions.

Long-term Context Chapter 1 of this volume deals with the long-term tectonic history of the Mediterranean basin, and this section therefore only highlights several particularly relevant aspects of change in the climatic and oceanographic setting. Overall, the Mediterranean is a relic ocean basin, representing the final stage of closure of the Tethys Ocean prior to continent–continent collision as the African plate converges with the Eurasian plate system. Note, however, that parts of the western basin are relatively young, and actively opening and deepening— notably the Tyrrhenian Sea. The proto-Mediterranean’s eastern connection with the open ocean (through the Levantine–Arabian region) closed roughly 18 Ma ago (Vergnaud-Grazzini 1985). Since that time, the only connection of the Mediterranean basin with the open world ocean has been through waterways in the west. Two such waterways connected the Mediterranean with the Atlantic Ocean: one across north Morocco (Rifian Strait) and one through the southern Iberian Peninsula (Betic Strait). Tectonic closure of the Betic and Rifian Straits led to massive evaporite deposition between 5.9 and 5.5 Ma, a phase known as the ‘Messinian Salinity Crisis’ (followed by the so-called ‘Lago Mare’ phase 5.5– 5.3 Ma) (Hilgen et al. 1995; Chapter 1). Re-establishment of open marine conditions following the Messinian Salinity Crisis appears to have been virtually synchronous everywhere in the Mediterranean basin, and may be ascribed to the tectonic opening of the Strait of Gibraltar. This event heralded the appearance of a basin that clearly began to approach the modern configuration. However, ongoing plate subduction processes (including slab detachment underneath southern Italy) caused continuing highly complex tectonic reshaping in the area. For example, the Tyrrhenian Sea underwent very rapid deepening and extension between ∼3 and ∼1.5 Ma, while tremendous uplift in southern Italy and parts of Greece has caused Late Pliocene/Early Pleistocene coastal sediments to be displaced to many hundreds of metres above modern sea level. The basin’s geological history can directly affect modern processes. When faults expose parts of the massive Messinian evaporite deposits to sea water in the basin, salt dissolution affects modern bottom-water properties.

The Marine Environment: Present and Past

This happens in the so-called ‘brine basins’ of the eastern Mediterranean. Dissolved salts in the bottom waters of these isolated depressions cause extremely high salinities, separated from the normal deep waters by a very sharp salinity gradient (halocline), which defines a strong density gradient (pycnocline). The oceanography and chemistry of brine basins are entirely different than in the open waters around them. Because of the extreme density stratification, the brines are not ventilated, and thus have become entirely oxygen-depleted. Sediments in these basins are often disturbed by masstransport processes, but on rare occasions undisturbed sections yield beautifully laminated cores, reflecting the fact that there is no benthic life to homogenize the sediments through bioturbation (among many others: Jongsma et al. 1983; Scientific Staff Cruise BAN84 1985; Troelstra et al. 1987; MEDRIFF consortium 1995; Wallmann et al. 1997). Major ‘global’ climate developments also need to be considered when studying palaeoclimatic and palaeoceanographic signals in the Mediterranean. The development towards a ‘glacial mode’ in the Northern Hemisphere started around 3.2–3.1 Ma (Shackleton and Opdyke 1977; Thunell and Williams 1983; Prell 1984). The Mediterranean environment was substantially affected by the Northern Hemisphere glaciations (Vergnaud-Grazzini 1985; Thunell et al. 1987, 1991). Ruddiman et al. (1987) found the first clear evidence for ice-rafting in the North Atlantic around 2.55 Ma, and Zachariasse and Spaak (1983) demonstrated that biogeographic patterns similar to the present originated around that time in the Mediterranean and adjacent Atlantic. The early development of Northern Hemisphere glaciation was associated with climatic change over the Mediterranean basin, characterized by increasing seasonal contrasts with very dry summers (Suc 1984; see also Thunell 1986). Suc (1984) argued that the ‘modern’ conditions with cool wet winters and hot dry summers first developed around 3.2 Ma, and that summer drought became more persistent after 2.8 Ma. Global atmospheric circulation modelling by Ruddiman and Kutzbach (1989) suggests that these developments may have resulted from northern hemispheric climate reorganization due to uplift of the Tibetan plateau, while the periodical appearance of steppe vegetation in the Mediterranean realm since 2.3 Ma (Suc 1984) would be related to large-scale expansions of Northern Hemisphere ice-sheets. The early glacial cycles had a mean periodicity of 41,000 years (obliquity forcing), which changed to a predominant periodicity of 100,000 years (eccentricity forcing) after the so-called

37

Mid-Pleistocene Transition, roughly 1.0 to 0.9 Ma (Shackleton and Opdyke 1973, 1976; Pisias and Moore 1981; Ruddiman et al. 1986, 1989). This change is well represented in Mediterranean isotopic, floral, and faunal records (e.g. Zachariasse et al. 1989, 1990; Lourens et al. 1992; Vergnaud-Grazzini et al. 1993; Chapter 4).

Modern Climate and Oceanography Climate The classical Mediterranean climate, discussed in detail in the following chapter, is characterized by warm and dry summers, and mild and wet winters. As such, it appears opposite to monsoon climates, which instead comprise a pluvial maximum in the warm months. The Mediterranean climate regime is due to the basin’s location on the transition between the climate conditions of the temperate westerlies that dominate over central and northern parts of Europe, and the subtropical high pressure belt over North Africa (Figure 2.4) (Boucher 1975; Lolis et al. 2002; Chapter 3). In summer, the subtropical high pressure conditions are displaced to the north and most of the Mediterranean experiences drought, especially the south-eastern sector. Polar front depressions may still reach the western Mediterranean, but they only exceptionally penetrate the eastern Mediterranean (Rohling and Hilgen 1991). During winter, the subtropical conditions are displaced southward, and the (northern sector of) the Mediterranean comes under the influence of the temperate westerlies with the associated Atlantic depressions that track eastward over Europe. Polar and continental air masses over Europe are channelled towards the Mediterranean through valleys between the mountainous topography of the northern Mediterranean margin. During winter and spring, intense cold and dry katabatic air flows are channelled through the lower Rhône Valley towards the Gulf of Lions (‘Mistral’), and also over the Adriatic and Aegean Seas (‘Bora’ and ‘Vardar’), causing strong evaporation and cooling of the sea surface (e.g. Leaman and Schott 1991; Saaroni et al. 1996; Poulos et al. 1997; Maheras et al. 1999; Casford et al. 2003; and references therein). Conditions for northerly air flow into the western and eastern Mediterranean are determined by interaction between an intense low over the central or eastern Mediterranean, and north-eastward extension of the Azores High (over Iberia, France, and

38

Eelco Rohling et al.

North Hadley Cell 23°N

L

ITCZ 0°

Tropical easterly jet South Hadley Cell

H 23°S

Ferrell Cell

Fig. 2.4. Northern Hemisphere summer atmospheric circulation pattern. The main winds are indicated as arrows. ITCZ = Inter-Tropical Convergence Zone, H = areas of high sea-level pressure, L = areas of low sea-level pressure (modified from Rossignol-Strick 1985; Reichart 1997).

southern Britain) or westward ridging of the Siberian High towards north-west Europe and southern Scandinavia (Maheras et al. 1999; Lolis et al. 2002). The wintertime low surface pressure conditions over the Mediterranean are a direct consequence of the high sea-surface temperatures due to the high thermal capacity of the basin’s watermasses (Lolis et al. 2002). The most pronounced basin-wide cold winter events (complementing widespread cold conditions over Europe) develop in association with positive sea-level pressure anomalies to the west or north-west of the British Isles and particularly low pressure over the Mediterranean, a configuration that reflects an extreme phase of the North Atlantic Oscillation (NAO) (Moses et al. 1987; Maheras et al. 1999). The main mode of climate variability in the Mediterranean is expressed by the so-called Mediterranean Oscillation (MO), a west–east pressure see-saw that is apparent both at the surface and at 500 hPa, especially in winter and spring (Maheras et al. 1999; Lolis et al. 2002). Statistical correlation has been found between the MO and the pressure see-saw of the NAO, where the low NAO index phase is associated with wet conditions in the western Mediterranean (Maheras et al. 1999; Lolis et al. 2002; Dünkeloh and Jacobeit 2003; and references therein). This confirms previous observations of direct NAO impacts on the western

Mediterranean, but for the eastern basin the relationship remains weakly established, except via dependence of the MO on the NAO (Dünkeloh and Jacobeit 2003). The statistically second main mode of winter variability, with important impacts on cyclogenesis in the basin and consequent precipitation in the north-eastern and southcentral parts of the Mediterranean, is the Mediterranean Meridional Circulation (MMC) (Dünkeloh and Jacobeit 2003). Cold and relatively dry northerly (meridional) airflow over warm sea surfaces causes intense cyclogenesis (formation of new depressions) in the northern sectors of the Mediterranean. Most cyclones observed in the Mediterranean are thus formed over the basin itself, although some Atlantic depressions may enter the (western) basin (Rumney 1968; Trigo et al. 1999; Chapter 3). Throughout the basin, however, winter cyclones are clearly linked to North Atlantic synoptic systems, as secondary lows when Atlantic systems interact with the Alps with additional cyclogenesis over the basin itself (Trigo et al. 2000). Cyclogenesis is most frequent in the western Mediterranean, especially over the Gulf of Genoa and Ligurian Sea, but the Aegean Sea is a major centre for winter-time cyclogenesis as well (Trewartha 1966; Rumney 1968; Boucher 1975; Cantu 1977; Trigo et al. 1999; Chapter 18). Most of the Genoan depressions track south-eastward down the coast of Italy

The Marine Environment: Present and Past

and then eastward or north-eastward across the Aegean Sea or northern Levantine seas (Trewartha 1966; Rumney 1968; Trigo et al. 1999; Lolis et al. 2002). Along the way, these depressions as well as those developing over other centres of cyclogenesis cause the winter precipitation that is so typical for the modern Mediterranean climate. The stable hydrogen and oxygen isotope composition of this precipitation follows a Mediterraneanspecific mixing line (the Mediterranean Meteoric Water Line, MMWL), which is different from the global Meteoric Water Line (MWL) due to the dominant contribution of moisture evaporated from the Mediterranean Sea into low-humidity air masses (Matthews et al. 2000; and references therein). Summer rainfall is low today, especially in the eastern basin. Although cyclogenesis occurs around Cyprus and the Middle East in summer, as a semi-permanent extension of the Indian monsoon low, dry summer conditions prevail as a consequence of adiabatic descent in the upper troposphere that is related to the intense Asian summer monsoon (Rodwell and Hoskins 1996; Trigo et al. 1999). Mean annual precipitation along the Mediterranean ranges from less than 120 mm in North Africa, to over 2,000 mm in portions of south-west Turkey and in the eastern Adriatic Sea along the slopes of the Dinaric Alps (Naval Oceanography Command 1987). Total evaporation in the entire Mediterranean increases towards the east, with an average of 1,450 mm y−1 (Malanotte-Rizzoli and Bergamasco 1991) to 1,570 mm y−1 (Béthoux and Gentili 1994). Strong rates of evaporation occur in areas subjected to strong winds, such as the Gulf of Lions and Ligurian Sea, the Aegean and Cretan Seas, and the southern part of the Turkish coast (MEDOC Group 1970; Miller 1974). Evaporation is weakest along the Moroccan and Algerian coasts (The Alboran Sea) where the air masses generally arrive from the Atlantic with relatively high air humidity. The basin-wide mean Mediterranean excess of evaporation over freshwater input [E (evaporation) – P (precipitation) – R (runoff)] has been variously estimated at ∼1,000 mm y−1 (Béthoux et al. 1999), 750 mm y−1 (Gilman and Garrett 1994), and 560–660 mm y−1 (Bryden and Kinder 1991). There is marked spatial variation in regional values (Chapters 3 and 8). Northern areas such as the Gulf of Lions, Adriatic and Aegean Seas show relatively low excess evaporation rates due to high freshwater inputs from the Rhône and Ebro rivers, the Po River, and the Black Sea, respectively. Southern regions show very high excess evaporation rates, especially in the eastern Mediterranean (Béthoux and Gentili 1994). The strong overall excess

39

evaporation results in a pronounced surface water salinity increase from west to east (MEDATLAS 1997) (Figure 2.2). Sea surface temperature values in the Mediterranean reflect a balance dominated by high energy gain from solar irradiation during the widespread subtropical high-pressure (clear) conditions in summer, and considerable (latent) heat loss during evaporation. As a result, sea surface temperature values increase towards the east and south throughout the Mediterranean. Winter values are around 10◦ C in the north-western Mediterranean and 15◦ C in the south-eastern Mediterranean, while summer values are around 21◦ C in the northwestern Mediterranean and 26◦ C in the south-eastern Mediterranean (Naval Oceanography Command 1987). The warmest season centres on July–August and the coldest on February–March. One further climate impact on the Mediterranean Sea must be mentioned. It concerns a ‘remote’ influence by a climate system that does not itself penetrate into the basin, namely the African monsoon. It used to influence the Mediterranean mainly through Nile River discharge, but has been severely curtailed since completion of the first stage of the Aswan High Dam in 1964. Prior to the anthropogenic control of the Nile, its average discharge was 8.4 × 1010 m3 yr−1 (4.5 × 1010 m3 yr−1 in lowflood years to 15.0 × 1010 m3 yr−1 in high-flood years), which from the mid-1960s has dwindled to a negligible amount (Nof 1979; Said 1981; Wahbi and Bishara 1981; Béthoux 1984; Rohling and Bryden 1992). Note that the reported discharge values illustrate that, even in the instrumental era, there was strong (threefold) interannual variability between high and low discharge years, which was mainly related to variability in the monsoon-fed contribution of the Blue Nile and Atbara rivers (see below). The Nile River comprises two different systems: the White Nile, which drains the equatorial uplands of Uganda in a regular, permanent manner; and the Blue Nile and Atbara, which receive highly seasonal African monsoon precipitation from the Ethiopian highlands. Nile hydrology has been summarized by Adamson et al. (1980) and Williams et al. (2000). In summary, these authors find that prior to extensive anthropogenic intervention (damming), a maximum of 30 per cent of the annual discharge of the Nile originated from the White Nile, and a minimum of 70 per cent from the Blue Nile and Atbara basins. The winter flow was dominated (83%) by the steady White Nile contribution, whereas the Blue Nile/Atbara component provides 90 per cent of the flow in summer. This seasonal contrast results from a massive increase in the Blue Nile and

40

Eelco Rohling et al.

TABLE 2.1. Contributions to total Nile discharge from the main tributaries Flood season (between August and October). Values in 106 m3 day−1

Regular flowing of water (outside flood seasons). Values in 106 m3 day−1

White Nile (from Lake Victoria = equatorial highlands) Blue Nile (Ethiopia) Atbara (Ethiopia)

70.0

37.5

485.0 157.0

7.5 0.0

Total

712.0

45.0

Sources: Hurst (1944); Said (1981).

Atbara discharge between a winter low and summer high (see also Table 2.1), with the monsoon-related peak occurring in the months August–October. The White Nile discharge shows a much smaller ratio of change between its annual peak and lowest monthly value (Table 2.1), and is highest between late September and January. Table 2.1 illustrates historical discharge values for the three main tributaries after Hurst (1944) and Said (1981) (according to those observations, the White Nile contribution to total annual discharge amounts to only 14%). The total suspended sediment load transport to the Mediterranean coast before closure of the Nile by the Aswan High Dam exceeded 1.0 × 108 tonnes yr−1 (Sharaf el Din 1977; El Dardir 1994; Stanley 1996). Since completion of the Aswan High Dam, there has been negligible Nile discharge and sediment transport into the Mediterranean through the Rosetta and Damietta outlets (UNDP/UNESCO 1978). Instead, salt water entering the mouth of the Rosetta extends some 25 km upstream to the Nile barrage at Mutubis. A little fresh water reaches the Mediterranean through the Manzalla, Burullus, and Idku lagoon outlets, and by pumping of lake Maryut water to the sea at Alexanderia (Stanley and Wingerath 1996). The suspended load that bypasses the Nile Delta to the shelf via Nile distributaries, lagoon outlets, and canals is about 15 per cent of the original (predam) load (Stanley et al. 1998). Apart from damming, the freshwater flow and sediment flux into the Mediterranean Sea were also curtailed due to the extensive irrigation network of canals and drains covering the entire Nile delta. Prior to its anthropogenic reduction, the Nile plume used to be distinctly traceable with the prevailing surface circulation in the easternmost Mediterranean, from the Nile delta east- and northwards along the eastern Levantine coast. It caused a zone with notably reduced

surface-water salinities and enhanced turbidity (suspended matter) (Reiss et al. 1999).

Surface-water Circulation Circulation in the Mediterranean Sea is driven by wind stress and thermohaline forcing (Robinson et al. 1992). Atlantic water (AW) enters the Mediterranean Sea as a surface flow through the Strait of Gibraltar, compensating for the net evaporative loss from the basin and the subsurface outflow. AW enters with a salinity of about 36.2 psu and temperature of about 15◦ C (Béthoux and Gentili 1994). As it migrates through the Strait of Gibraltar, AW mixes with upwelled Mediterranean Intermediate Water (MIW), creating Modified Atlantic Water (MAW) which has higher temperatures (16◦ C) and salinities (36.5%) (La Violette 1986; Tintoré et al. 1988; Arnone et al. 1990; Heburn and La Violette 1990). In the Alboran Sea, MAW is present along the southern Spanish coast as a strong jet (with speeds up to several kilometres an hour) approximately 20 km wide and extending to a depth of 150 m (Pistek et al. 1985). The strength of the jet initiates the formation of two anticyclonic gyres (Figure 2.5), the positions of which fluctuate on timescales of 3–4 weeks (Heburn and La Violette 1990). As the MAW flows eastward along the Spanish coast to Almeria, it converges with resident Mediterranean waters. The subsequent deflection of MAW towards Oran on the Algerian coast forms a well-defined frontal zone along the eastern edge of the Eastern Alboran Gyre (Figure 2.5). This front extends to a depth of 200 m and has a width of approximately 35 km (Cheney and Doblar 1982). The Almeria–Oran Front, as it is known, is thought to be a permanent feature, although its position and intensity are controlled by the degree of development of the Eastern Alboran Gyre (Tintoré et al. 1988). To the east of the Alboran Sea, MAW is concentrated along the northern coast of Africa in the Algerian Current. To its north, northward branches of the MAW form part of various larger-scale cyclonic gyres (Figure 2.3), while smaller anticyclonic gyres are found to the south of the Algerian Current. Waters flowing northwards on both sides of Corsica, the western and eastern Corsica currents, join and form the northern cyclonic gyres in the Gulf of Lions, where the Mistral winds in winter initiate a series of processes leading to the formation of Western Mediterranean Deep Water (WMDW) (e.g. MEDOC Group 1970; Gascard 1978; Leaman and Schott 1991; Robinson and

The Marine Environment: Present and Past

41

39°N

SPAIN

Almeria

Almeria-Oran Front

Gibraltar West Alboran Gyre

East Alboran Gyre

Al

37°N

Algerian Basin

r ge

ian

Current

Algiers

Oran

ALGERIA 35°N 6°W

4°W

2°W

0°

2°E

Fig. 2.5. Schematic illustration of surface circulation in the Alboran Sea (modified from Tintoré et al. 1988).

Golnaraghi 1994; Rohling et al. 1998b; and references therein). MAW enters the eastern Mediterranean through the Strait of Sicily with salinities between 37.0 and 38.5 psu (38.5 is the salinity at the flow reversal boundary; Garzoli and Maillard 1979). It feeds the Ionian Current and Mid-Mediterranean Jet (MMJ) through the Ionian Sea and Levantine basin, respectively. The MMJ bifurcates several times to form a series of cyclonic and

anticyclonic gyres interconnected by jets flowing at speeds of 20–30 cm s−1 (Robinson et al. 1992) (Figures 2.3 and 2.6). One branch of the MidMediterranean Jet flows to Cyprus and then north- and westwards to become the Asia Minor Current (Figure 2.3). It must be noted that salinity values of MAW increase steadily as it travels from west to east, due to continued evaporation (Wüst 1961; Malanotte-Rizzoli and Hecht 1988).

Fig. 2.6. Schematic illustration of the main gyres associated with Atlantic surface flow. IAS = Ionian–Atlantic Stream, CC = Cilician Current, AMC = Asia Minor Current, MMJ = Mid-Mediterranean Jet (modified from Robinson et al. 1992).

42

Eelco Rohling et al.

Intermediate-water Circulation

38.5

39.5

500

1000

Depth (m)

During winter, surface waters in the Levantine basin experience enhanced mixing and evaporation as a consequence of strong winds associated with cold, dry air masses tracking the eastern Mediterranean at this time of year (Ozsoy 1981), especially in the Cyprus– Rhodes area. A subsequent combination of low temperatures (15–16◦ C) and high salinities (∼39.2 psu, with extremes to 39.5 psu; Wüst 1960) in surface waters creates favourable conditions for vertical convection and the consequent formation of Levantine Intermediate Water (LIW). This watermass is characterized by a salinity maximum, and spreads throughout the eastern and western Mediterranean basins, forming the major constituent of the Mediterranean Intermediate Water (MIW) (Figure 2.2). It resides between 150 and 600 m water depth, and its transition to surface MAW is marked by a distinct salinity gradient, or halocline (Figures 2.2 and 2.7). There is no comparable source region for intermediate water formation in the western Mediterranean basin. From its source area, LIW/MIW flows westwards, penetrating the Ionian and Adriatic Seas. On approaching the Strait of Sicily, part of the subsurface watermass is recirculated within the eastern basin, while the remainder continues to enter the western Mediterranean basin. The actual ratio of recirculation to efflux remains to be established (Robinson et al. 1992). At the Strait of Sicily, MIW remains distinctive within the water column, although with somewhat reduced temperature (14◦ C) and salinity (38.75%) values compared to those in the LIW source area, due to later admixtures (Garzoli and Maillard 1979). On leaving the Strait, MIW settles between about 200 and 600 m, and splits into main branches that go: into the Tyrrhenian basin; along the western side of Sardinia; and along the AlgerianMoroccan coastlines to exit through the Strait of Gibraltar from the Mediterranean into the North Atlantic. MIW enters the Alboran Sea at a depth between 200 and 600 m, with temperatures and salinities of 13.2◦ C and 38.5% respectively, flowing in a westward direction towards the Strait of Gibraltar at velocities of 1–2 cm s−1 (Parrilla et al. 1986; Richez and Gascard 1986). Since the subsurface outflow through the Strait of Gibraltar displays temperature and salinity values of about 13◦ C and 38.2–38.4 psu, compared to values of 15–16◦ C and 36.1–36.2 psu in the surface (AW) inflow (e.g. Wüst 1960, 1961; Gascard and Richez 1985; Bryden et al. 1994), it is obvious that the Mediterranean experiences both net evaporation and net cooling (Garrett 1994). The subsurface Mediterranean Outflow has

Salinity (p.s.u.) 37.5 0

1500

2000

2500

3000

Fig. 2.7. Typical salinity profiles for the western (dots) and eastern (squares) Mediterranean basins (modified from Rohling and Bryden 1992).

a flux in the order of 1 Sv (Bryden and Kinder 1991) to 1.5 Sv (Béthoux and Gentili 1994) (1 Sverdrup = 1 × 106 m3 s−1 ). It settles between 1,000 and 1,500 m depth in the North Atlantic Ocean (e.g. Wüst 1960; Stommel et al. 1973; Reid 1979; Price et al. 1993).

Deep water Circulation The western and eastern Mediterranean basins each have their own source of deep water, which settles below the MIW. Western Mediterranean Deep Water (WMDW) is formed in the north-west Mediterranean, particularly the Gulf of Lions, and Eastern Mediterranean Deep Water (EMDW) in the Adriatic and Aegean Seas. Today, there is such consistent deep water ventilation from these regions that both the western and eastern Mediterranean are characterized by well-oxygenated deep and bottom waters, with oxygen concentrations

The Marine Environment: Present and Past

typically varying in reported ranges of 4.0–4.7 ml l−1 or 180–210 Ï mol kg−1 (Wüst 1960; McGill 1961; Miller et al. 1970; Schlitzer et al. 1991; Klein et al. 1999; Roether and Well 2001). The following two sections discuss the mechanisms for WMDW and EMDW formation in more detail.

Western Mediterranean Deep Water (WMDW) The Gulf of Lions is the key area for the western Mediterranean deep circulation. The surface circulation in this area is characterized by a distinct cyclonic gyre (MEDOC-group 1970) (Figure 2.3). In winter (January/ February), cold and relatively dry Mistral winds over this region initiate WMDW formation. Three phases can be distinguished: (1) the preconditioning phase, (2) the violent mixing phase; and (3) the sinking and spreading phase (Figure 2.8) (MEDOC Group 1970). During the preconditioning phase, a reduction occurs in the stability of the water column due to winter cooling that leaves surface waters with low temperatures (10–12◦ C), high salinities (38.40 psu), and consequently elevated densities (Wüst 1961; MEDOC Group 1970; Leaman and Schott 1991). At this time mixing occurs in the surface waters but the vertical profile still remains a three-layered one: (1) a relatively fresh and cold surface layer, (2) a warm saline intermediate layer, and (3) a cold and medium-saline deep layer (Figure 2.8). The onset of strong north-westerly Mistral winds (MEDOC-group 1970) initiates an intensification of the basin’s cyclonic circulation, which causes a shallowing of the pycnocline from a usual depth of approximately 200–250 m (Perkins and Pistek 1990) to <100 m (see Rohling et al. 1995). The preconditioning phase is followed by a phase of violent mixing. Throughout February, the density of surface waters increases due to cooling and intense evaporation (2 cm day−1 ; MEDOC-group 1970), eliminating the gradient between the surface and intermediate waters. This results in ‘chimneys’ of convective mixing that reach throughout the water column to great depths (>2000 m), developing within the centre of the gyre (MEDOC-group 1970; Leaman and Schott 1991). Incidentally, the existence of discrete ‘chimneys’ of deep convective mixing was observed for the first time in this area during the MEDOC study, and similar features have since been recognized in other areas of deep water formation (notably the Norwegian Sea). The geographical extent of the region of deep water formation is characterized at the surface by high salinities (38.4 psu) mixed up from below, from the intermediate water (MEDOC-group 1970).

Preconditioning phase

43

E E

Surface water

Mediterranean intermediate water Western Mediterranean cool, deep water Violent mixing and deep convection phases

E

Surface water

Mediterranean intermediate water Western Mediterranean cool, deep water Fig. 2.8. Schematic illustration of the preconditioning phase, and the violent mixing and deep convection phase. E = Evaporation (modified from Rohling et al. 1998b).

Then follows a phase of sinking and spreading. As the stormy period ceases, the mixed water sinks rapidly to form WMDW (Figure 2.8). The newly formed watermass is characterized by a relatively high oxygen content (4.4–4.7 ml l−1 ), and spreads horizontally between 1,500 and 3,000 m into the Balearic basin and Tyrrhenian Sea (Wüst 1961). On entering the Alboran Sea, WMDW forms a narrow (∼20 km) boundary current flowing westward along the Moroccan coast before entering the Strait of Gibraltar. In the Alboran Sea WMDW reaches speeds of approximately 5 cm s−1 , and it contributes an estimated 0.3 Sv (25%) to the outflow over the Gibraltar Sill (Parrilla et al. 1986; Richez and Gascard 1986).

44

Eelco Rohling et al.

Eastern Mediterranean Deep Water (EMDW) Throughout the period of oceanographic observation, until the late 1980s/early 1990s, the Adriatic Sea was found to be the main source area of EMDW formation (Pollak 1951; Wüst 1961; Malanotte-Rizzoli and Hecht 1988; Robinson et al. 1992). In winter, cold and dry north-easterly winds (Bora) cause intense cooling of the North Adriatic shelf waters (Ozsoy 1981), which are of relatively low salinity due to dilution with fresh water from the Po river. The resultant cold waters flow towards the deep south Adriatic basin, where mixing occurs with the warmer but more saline MIW that penetrates the South Adriatic across the Otranto Sill. The mixing of the cold and relatively low-salinity shelf waters with warm and highly saline MIW results in the formation of Adriatic Deep Water (ADW). Although ADW has a lower salinity (<38.7%) than MIW, it is also cooler, with values between 13.0 and 13.6◦ C. The resultant higher density of ADW allows it to settle below the MIW down to the greatest depths in the eastern Mediterranean basin (in any case, it used to do so until the late 1980s/early 1990s). It thus forms a major component of the EMDW. The EMDW circulates in a deep western boundary through the Ionian Sea before entering into the Levantine basin (Robinson et al. 1992). Roether and Schlitzer (1991) constructed a 22-box model for the deep water flow in the Ionian and Levantine basins. Their results indicate that the thermohaline circulation in the eastern Mediterranean at that time consisted of a single vertical cell through both basins, driven by Adriatic deep water formation. The derived rate of deep water supply from the Adriatic Sea into the eastern Mediterranean was 0.29 Sv and the turnover time about 126 years (Robinson et al. 1992). The importance of the Aegean Sea as a contributor to EMDW ventilation has been intensely debated. Pollak (1951) rejected the hypothesis that the Aegean Sea is a source of deep water formation, arguing that the Adriatic is the only source. Wüst (1961) disagreed, stating that the Aegean Sea source may be minor but not negligible. Indeed, Miller (1963) reported evidence that Aegean Deep Water (AeDW), contributing to the EMDW, formed sporadically in the Aegean Sea before flowing into the Levantine basin via the Straits of Kasos and Karpathos. Roether et al. (1983), however, concluded from 3 H and 3 He data that the bottom waters of the eastern Mediterranean were formed exclusively in the Adriatic. The entire debate took a dramatic turn when observations obtained on RV Meteor cruise M31-1 (January–February 1995) reported that an influx of Aegean Sea water had replaced approximately 2 per cent of the deep and bottom waters

of the eastern Mediterranean, strongly enhancing the observed deep/bottom-water salinities and displacing older waters upwards (Roether et al. 1996). It was inferred that Aegean Sea outflow now contributed up to 65 per cent to the deep and bottom waters of the eastern Mediterranean (ibid.). Circulation in the Aegean Sea is mainly controlled by the regional climate, local riverine inputs that occur mainly in winter, and the Black Sea surface-water outflow that increases in summer (Poulos et al. 1997). Annual surface temperatures in the Aegean Sea vary from <13◦ C in winter to >24◦ C in summer. Salinity varies from <31.0 to >39.0 psu (Poulos et al. 1997), with locally lowest values (26 psu) in summer as a result of the Black Sea outflow (Yüce 1995). The formation of AeDW is again closely related to the influences of salty Levantine Intermediate Water. As the LIW-derived Aegean Intermediate Water (AeIW) travels north along the Turkish coast, the prevailing offshore winds cause its upwelling to the surface (Lascaratos 1989; Yüce 1995). In these shallow eastern shelf areas, the AeIW consequently forms a single uniform watermass from the surface to the sea floor. As the upwelled AeIW progresses northwards at the surface, it is directly affected by the regional climate. Wintertime cold and dry northerly outbreaks of polar/continental air masses cause strong surface buoyancy loss from the Aegean Sea, through cooling and increasing salinities (Theocharis and Georgopoulos 1993). This leads to the formation of AeDW, which today fills the Aegean basin below 300 m (Bruce and Charnock 1965; Burman and Oren 1969; Miller et al. 1970; Miller 1972; Theocharis 1989; Yüce 1995). As stated before, AeDW formation was long considered of minor importance to the deep water ventilation of the open eastern Mediterranean, but recent studies show that specific (cold) climatic forcing over the Aegean in the early 1990s caused higher-salinity AeDW to replace Adriatic Deep Water (ADW) as the dominant deep water in the open eastern Mediterranean (Roether et al. 1996, Samuel et al. 1999; Klein et al. 1999). This event, which has been named the ‘Mediterranean Transient’ initiated a new mode of deep ventilation in the eastern Mediterranean basin that has persisted until today (B. Klein, pers. comm. Nice, April 2004).

Quaternary Climatic and Hydrographic Changes The Quaternary Mediterranean palaeoclimatic and palaeoceanographic history reveals marked variability

The Marine Environment: Present and Past

with both orbital (‘Milankovitch’) and so-called ‘sub-orbital’ or ‘sub-Milankovitch’ periods. The former refers to variability with periods similar to those of the astronomical cycles of eccentricity, obliquity, and precession. The latter refers to variability at periods shorter than those of the astronomical cycles (i.e. shorter than 19,000 years). The astronomical cycles (eccentricity with periods of ∼400 and ∼100 ka, obliquity with a main period of 41 ka, and precession with periods of 23 and 19 ka) govern changes in the intensity and distribution of insolation. They have especially important impacts on glacial–interglacial alternations and monsoon intensity. The Pleistocene glacial–interglacial cycles followed first obliquity and then (since ∼900 ka BP) eccentricity time scales. Typical glacial–interglacial contrasts in the Mediterranean’s climatic and oceanographic features are well illustrated by a comparison of the Last Glacial Maximum with the current and previous interglacial maxima. Superimposed on the glacial–interglacial cycles, frequent episodes of organic-rich sediment accumulation have occurred in the (eastern) Mediterranean, timed according to monsoon maxima as determined by the eccentricity-modulated precession cycle. Sub-Milankovitch climate variability is globally widespread as well, but its origin remains elusive. The millennial-scale ‘Dansgaard-Oeschger events’ were first observed in temperature proxy records from the welldated Greenland ice cores (Langway et al. 1985; Dansgaard et al. 1993; Grootes et al. 1993), and now provide a widely accepted template for sub-Milankovitch variability of the last 110 ka in the North Atlantic– Eurasian region (e.g. Broecker 2000; Voelker et al. 2002; Rohling et al. 2003; Hemming 2004; and references therein). Notable ‘tie-points’ are associated with the extreme cold events known as ‘Heinrich events’, when massive ice-berg flotillas caused a great pulse of icerafted debris deposition and melt-water flooding in the North Atlantic, with cold and arid climatic effects that were noted on an at least Northern Hemispheric scale (see overviews in Rohling et al. 2003; Hemming 2004). The Mediterranean sedimentary record contains ample evidence of millennial to centennial timescale sub-Milankovitch variability that has been convincingly related to this template. In the following sections, we will first introduce relevant aspects of orbital forcing, followed by overviews of: (1) the character of glacial-interglacial cycles in the Mediterranean; (2) the impact of insolation-driven monsoon maxima on Mediterranean hydrography and sedimentation; and (3) the expressions of centennial- to millennial-scale variability in the basin.

45

The Orbital Periods and Insolation at 65 ◦ N In the 1860s to 1870s, James Croll pioneered an astronomical theory of climate change. In the late 1930s, the Serbian engineer Milutin Milankovitch expanded this theory, calculating the astronomically determined fluctuations in the intensity and distribution of solar radiation onto the earth, presented in the form of insolation reconstructions for various discrete latitude bands. This section briefly introduces the astronomical cycles, starting with eccentricity and precession, and concluding with the obliquity or tilt cycle. First, it is useful to summarize the nature of the cardinal points in the seasonal cycle. On an annual timescale, the position of the earth’s rotational axis, tilted relative to the plane of the earth’s orbit around the sun, is fixed in space. Today, the North Pole points towards the star Polaris. Northern Hemisphere winter starts with the Northern Hemisphere (‘boreal’) winter solstice, when the North Pole lists directly away from the sun, resulting in the shortest day on the Northern Hemisphere. Next, the boreal spring (‘vernal’) equinox marks the start of boreal spring. During an equinox, the boundary between the illuminated and dark half-globes passes through both Poles, so that day and night have identical durations at all points of the world. Then follows the boreal summer solstice, when the North Pole lists directly towards the sun, resulting in the longest day on the Northern Hemisphere. It marks the start of boreal summer. Thereafter, the boreal autumnal equinox is reached, which marks the start of boreal autumn. Eccentricity concerns the shape of the earth’s orbit around the sun, which varies from near circular to distinctly elliptical. An ellipse has two focal points, and as the ellipse transforms to a circle, the two focal points approach one another. Eccentricity is expressed as a measure of the distance between the two focal points relative to the distance along the long axis of the ellipse. The eccentricity of the earth’s orbit varies between almost 0 and about 6 per cent. The sun occupies one of the focal points of the earth’s orbit, the other one is empty. The non-circular shape of the orbit dictates that earth passes a point nearest the sun (‘perihelion’) and a point furthest away from the sun (‘aphelion’) during each of its annual revolutions around the sun. Today, perihelion occurs close to the boreal winter solstice, and aphelion close to the boreal summer solstice. Note that, when the orbit is near circular—an eccentricity minimum—the earth’s distance to the sun is virtually constant through the year. The eccentricity of the earth’s orbit changes in a cyclic fashion, with three main periods: 94,800 years,

46

Eelco Rohling et al.

123,800 years, and 404,000 years. Palaeoclimate studies commonly approximate these with apparent periods of 100,000 and 400,000 years. The impact of eccentricity on insolation is primarily through modulation of the effects of precession. Precession refers to the fact that the earth’s rotational axis relative to its orbital plane is not fixed in space, but displays a long-term wobble, similar to the axis of a spinning top. This changes the direction of the axis in space, so that the earth’s North Pole, which today points towards Polaris, points towards Vega after half a precession cycle, and back towards Polaris again after a complete precession cycle. A full precession cycle takes 26,000 years, but due to other complications in the earth–sun motions (the entire earth orbit itself slowly rotates around the sun about once for every four precession periods) the precession cycle manifests itself in insolation with two dominant periods: one around 23,000 years and the other around 19,000 years. Precession causes a very slow shifting of the dates of the solstices and equinoxes along the orbit. A quarter of a cycle ago (about 5,500 years BP), therefore, perihelion occurred near to the boreal autumnal equinox. Half a cycle ago (about 11,000 years BP), perihelion occurred close to the boreal summer solstice. Three quarters of a cycle ago (about 16,500 years BP), perihelion coincided with the boreal vernal equinox, and a full cycle ago the situation concerning precession was similar to the present. The climatic impacts of the precession and eccentricity cycles need to be viewed together. Today, in its slightly elliptical orbit, the earth is at perihelion around the boreal winter solstice (3 January and 21 December, respectively). It is at aphelion around the boreal summer solstice (4 July and 21 June, respectively). When the orbit approaches a circle, these distance differences would have negligible effects. However, since some eccentricity applies, the solar radiation on illuminated places of the globe will be somewhat more intense in boreal winter (austral summer) than in boreal summer (austral winter). This weakens the seasonal contrast on the Northern Hemisphere, and strengthens it on the Southern Hemisphere. The precession cycle then shifts the distribution of the seasons around the elliptical orbit. Half a precession cycle ago, perihelion occurred near the boreal summer solstice and aphelion around the boreal winter solstice, which enhanced the seasonal contrast on the Northern Hemisphere. The cycle of obliquity concerns changes in the angle of the earth’s rotation axis relative to the perpendicular of the plane of the earth’s orbit over a period of 41,100

years, between 22.5 and 24.5 degrees. Today, the angle is about 23.5 degrees, so that the sun during the boreal summer solstice stands directly overhead at about latitude 23.5◦ North. This represents the maximum North latitude where the sun at any one time in the year reaches a directly overhead position—the Tropic of Cancer. During the boreal winter solstice (austral summer solstice) this condition is reached at about latitude 23.5◦ South—the Tropic of Capricorn. On a perfectly spherical earth, the obliquity cycle would therefore shift the position of the Tropics between latitudes 22.5 and 24.5◦ (the actual values are 22.04 and 24.45◦ ). In addition, the obliquity (or ‘tilt’) of the axis affects the amount of sunlight received at the high polar latitudes. For strong tilt, the poles receive more sunlight, and the sun’s rays also reach the polar surface at a higher angle, which both increases the energy received per unit area and decreases reflection. Variations in the astronomical parameters have now been reliably calculated back to 10–15 Ma or so (Laskar et al. 1993; Laskar 1999). Although many studies discuss palaeoclimatic records in terms of changes in the individual orbital parameters, most concentrate on their combined influence on insolation changes in specific latitude bands. Particular interest concerns the insolation changes at 65◦ North. At this latitude, the earth is occupied by great landmasses, which causes a sensitive setting for responses to insolation. Astronomically determined insolation records have been used as a template for glacial–interglacial variations reflected in marine stable oxygen isotope series (e.g. Imbrie et al. 1984, 1992; Martinson et al. 1987), and for variability in the Northern Hemisphere’s Indian/Asian and African monsoon systems through time (among others, Rossignol-Strick 1983, 1985; Hilgen 1991a , b; Hilgen et al. 1993, 1995; Lourens et al. 1992, 1996, 2001; Prell and Kutzbach 1987; Shimmield et al. 1990; Clemens and Prell 1990; Clemens et al. 1996). Because the insolation series is accurately calculated from astronomical changes with time, correlations of palaeoclimate records to the insolation records offer a sound insight into the chronology of the records. This concept lies at the heart of ‘astronomical or orbital tuning’ of the geological timescale. Great advances have been made in the development of an astronomically tuned geochronology, both on the basis of deep-sea stable oxygen isotope variations (e.g. Imbrie et al. 1984, 1992; Martinson et al. 1987; Pälike and Shackleton 2000), and on the basis of Mediterranean sapropel occurrences, other lithological alternations, and stable isotope data (e.g. Hilgen 1991a, b; Hilgen et al. 1993, 1995; Lourens et al. 1992, 1996, 2001).

The Marine Environment: Present and Past

Glacial Cycles in the Mediterranean The major glacial–interglacial climate oscillations of the Pleistocene had large impacts on the Mediterranean environment (Chapters 4 and 12). The development of a fixed anticyclone over the north European ice sheet and colder sea surface temperatures during glacial times are thought to have resulted in colder and drier conditions (i.e., reduced moisture supply from colder air masses) (Rognon 1987), with a likely increased seasonality of precipitation over the Mediterranean (Prentice et al. 1992). The present climate conditions are characterized by a high interglacial sea-level position, relatively dense vegetation cover, relatively high infiltration rates, and moderate river discharges, as summarized for the western Mediterranean by Rose et al. (1999). Glacial times, in contrast, were characterized by a low global sea level, open vegetation with large areas of bare ground and unconsolidated sediments, soils affected by high physical stresses, and highly peaked riverdischarge regimes (ibid.). Past vegetation (pollen) and lake-level records confirm the contrast between generally warm and relatively humid interglacial conditions and cold, relatively arid glacial conditions in the Mediterranean basin (e.g. Wijmstra et al. 1990; Digerfeldt et al. 2000; Elenga et al. 2000; Magri and Parra 2002; Tzedakis 1993, 1999; Chapters 4 and 9). Also, wind-blown dust transport into the Mediterranean was high during glacial times, suggesting enhanced aridity (Dinarès-Turell et al. 2003; Larrasoaña et al. 2003; Chapter 14). Glaciers and small ice-caps also developed in the high mountains of the Mediterranean during cold stages of the Pleistocene (Chapter 12 and Hughes et al. 2006), and the palaeoclimatic significance of these ice masses is discussed in Chapter 12. The most robust palaeoenvironmental characteristic of the last glacial maximum (LGM, ∼20, 000 ka BP) is a global sea-level low-stand at 120 or 125 m below the present-day level (Fairbanks 1989, 1990; Rohling et al. 1998a; Siddall et al. 2003; Peltier 2004). This lowering had serious impacts on Mediterranean hydrography, due to its effect on the hydraulically controlled exchange of watermasses through the Strait of Gibraltar. Any such sea-level lowering would cause a severe reduction in the exchange transport (to roughly half the modern value; Rohling and Bryden 1994; Rohling 1994; Myers et al. 1998; Matthiessen and Haines 2003). This would increase the residence-time of waters within the Mediterranean basin (calculated as Volume/Flux out), causing longer exposure of Mediterranean waters to the strong net evaporation, which leads to a considerable increase in salinity (stable oxygen isotope data

47

substantiate the impact of this concentration effect, as shown below) (Figure 2.9). Any subsequent climatic amelioration associated with sea-level rise would cause a rapid reduction in salinities within the basin, which would be conducive to poor deep water ventilation (Rohling 1994; Matthiessen and Haines 2003). The lower glacial sea-level position is also thought to have caused a shoaling of the density gradient (pycnocline) between intermediate and surface waters within the Mediterranean, with impacts on the plankton community structure by supporting a widespread deep chlorophyll maximum (Rohling and Gieskes 1989; Rohling 1991a ; Rohling and Bryden 1994; Myers et al. 1998). The aforementioned glacial concentration effect would furthermore have enhanced the salinity contrast between the Mediterranean and the open ocean, and so between outflow and inflow through the Strait of Gibraltar. Effectively, the Mediterranean outflow flux would have been reduced, but its density contrast with ambient Atlantic waters would have been significantly enhanced, resulting in a deeper-settling, smaller-volume glacial Mediterranean Outflow plume in the Atlantic (e.g. Rohling 1997). Similar impacts of sea-level lowering apply to other concentration basins, such as the Red Sea (e.g. Rohling and Zachariasse 1996; Rohling et al. 1998a ; Siddall et al. 2003; and references therein). The contrast between glacial and interglacial conditions in the Mediterranean is particularly obvious when comparing Sea Surface Temperature (SST) reconstructions for the LGM with those for the most recent, Holocene, Climate Optimum (HCO, 9–6 ka BP), the time when the most recent sapropel was deposited (Figure 2.10). The reconstructions are based on analysis of abundance data of planktonic foraminifera (Figure 2.11) from sediment cores with a recently developed artificial neural network method to derive Mediterranean temperature values (Hayes et al. 2005). On a basin-wide scale, these reconstructions corroborate the basic SST assumptions that were used previously in modelling studies (Myers et al. 1998). Complementary information on changes in the net freshwater budget between the LGM and HCO may be obtained from analysis of spatial patterns in stable oxygen isotope values through the basin (Kallel et al. 1997a , b; Rohling 1999a , b and references therein). The stable isotope data can be used to assess changes in salinity, subject to significant caveats (Rohling 1999a ; Rohling et al. 2004). Myers et al. (1998) used the isotope distributions to infer idealized LGM and HCO distributions of sea surface salinity (SSS) for use as restoring boundary

48 45°N

Eelco Rohling et al.

(a) 36.8 36.6

.6 36

40°N

36.2

36

.0

35.8

.8

36

36.0

37.

.8

37.0

36

35°N 36.6

0

36.8

30°N 10°W 45°N

0°

10°E

20°E

30°E

(b) 39.2

.0 39

40°N 38.4

.6 39

8

40

41.2

.0

40.

.2

39

.2

40

40

35°N

.4

30°N 10°W

0°

10°E

20°E

30°E

Fig. 2.9. Estimated sea surface salinity distribution for (a) the Holocene Climate Optimum and (b) the Last Glacial Maximum (modified from Myers et al. 1998), based on surface-water ‰18 O distribution patterns (De Rijk et al. 1999).

conditions in their modelling experiments of Mediterranean palaeocirculation. Overall, such studies suggest that net evaporation from the Mediterranean during the LGM may not have been very different from the present, suggesting that any reduced effective moisture supply from colder air masses was approximately offset by reduced evaporation from a colder sea. The higher LGM to present salinity contrast in the Mediterranean, relative to that in the open ocean, appears to be dominated by the glacial concentration effect (i.e. the impact of 120–125-m glacial sea-level lowering on

watermass exchange through the 284-m deep Strait of Gibraltar).

Monsoon Maxima and Mediterranean Sapropels Sapropels are dark, often laminated, organic-rich sediments found intercalated with the normal, organic-poor, hemipelagic sediments throughout the entire eastern Mediterranean (Figure 2.12). There are rare examples also from the western Mediterranean, especially

The Marine Environment: Present and Past

49

45°N

40°N

35°N

(a) Holocene Optimum 30°N 10°W

0°

10°E

20°E

30°E

°C 7+ 9+ 11+ 13+

45°N

15+ 17+ 19+ 21+

40°N

35°N

(b) Last Glacial Maximum 30°N 10°W

0°

10°E

20°E

30°E

Fig. 2.10. Annual sea surface temperature (SST) reconstructions for the Holocene Climate Optimum and Last Glacial Maximum (LGM) based on the artificial neural network (ANN) technique. To avoid any impacts of the 8.2 ka BP cold event, we defined the Holocene Climatic Optimum (HCO) as the interval between 8.3 ka BP and 9.5 ka BP. A total of 42 cores was selected for this time slice. In contrast, 37 cores were used for the LGM between 19 and 23 ka BP. The SST estimates for each core were obtained by calculating the average SSTs from all the samples within the defined time slices, based on a calibration of >300 core-top samples (modified from Hayes et al. 2005).

from Pliocene times. Sapropels range from a few millimetres to more than a metre in thickness, and have been deposited intermittently throughout the Neogene and Quaternary (among countless others: Kullenberg 1952; Olausson 1961; van Straaten

1972; Cita 1973; Cita et al. 1977; Vergnaud-Grazzini et al. 1977; Thunell et al. 1977, 1983; Stanley 1978; Williams et al. 1978; Cita and Grignani 1982; Rossignol-Strick et al. 1982; Rossignol-Strick 1983; Vergnaud-Grazzini 1985; Rohling and Gieskes 1989;

3

4

2

1

8 5

7

6

9

12

10

13

11

14

15

Fig. 2.11. Scanning electron microscope images of the carbonate shells of several planktonic foraminiferal species that live in the Mediterranean Sea. 1, 2 Globigerina bulloides d’Orbigny; 1 umbilical view and 2 spiral view. 3, 4 Globoturborotalita rubescens Hofker; 3 umbilical view and 4 spiral view. 5, 6 Turborotalita quinqueloba (Natland); 5 umbilical view and 6 spiral view. 7 umbilical view and 8 umbilical view. 9, 10 Globigerinoides sacculifer (Brady); 9 umbilical view and 10 spiral view. 11, 12 Globigerinella digitata (Brady); 11 umbilical view and 12 spiral view. 13–15 Globigerinella siphonifera (d’Orbigny); 13 umbilical view, 14 peripheral view, and 15 spiral view. Each scale bar represents 100 Ï m.

The Marine Environment: Present and Past

51

Fig. 2.12. Example of a laminated sapropel in a freshly opened sediment core. The thick dark bed recovered over two core sections represents sapropel S5 from the previous interglacial maximum, 124–119 thousand years ago. The core was recovered during cruise M53-1 of RV Meteor in November–December 2001 (chief scientist Prof. Ch. Hemleben).

Emeis et al. 1991, 1998, 2003; Hilgen 1991a, b; Hilgen et al. 1993, 1995; Rohling 1994; Van Os et al. 1994; Lourens et al. 1992, 1996, 2001; Nijenhuis et al. 1996; Rohling 1999b; Meyers and Negri 2003; and contributions and references therein). Sapropels are commonly marked by an absence of benthic foraminifera, and are preceded by a short interval containing benthic faunas indicative of severe bottom-water oxygen depletion (such faunas sometimes return within, or persist into/through, the sapropel) (e.g. van Straaten 1972; Nolet and Corliss 1990; Verhallen 1991; Rohling et al. 1993b 1997; Nijenhuis et al. 1996; Jorissen 1999; Mercone et al. 2001; Casford et al. 2003; Schmiedl et al. 2003). In marine cores, sapropels are recognizable as beds ranging in colour from dark grey to olive green and black. Exposed in land-sections, sapropels appear in notably darker shades of grey than surrounding beige to blue clays, and

commonly weather into distinct reddish-brown hues. Sapropels may display remarkably well-preserved lamination (Figure 2.12). The shallowest reported occurrence of the four youngest (most cored) sapropels in the open eastern Mediterranean is ∼300 m (Shaw and Evans 1984; Rohling and Gieskes 1989; Rohling et al. 1993a ). In the Adriatic Sea, the upper depth limit seems to have been at a deeper level, below 400 m (Jorissen et al. 1993). In the Aegean Sea, sapropels are found up to 120 m water depth (Perissoratis and Piper 1992; Casford et al. 2002). Following almost six decades of research on Mediterranean sapropels since their initial discovery in marine sediment cores recovered during the Swedish Deep-Sea Expedition of 1946–7, a general (but not unanimous) consensus has emerged that sapropels were formed during times with a combination of (1) enhanced abundances of organic matter sinking

52

Eelco Rohling et al.

and Grignani 1982; Rohling 1994; Emeis et al. 1998; Cramp and O’Sullivan 1999; Rohling et al. 2004). These impacts will be discussed below, and are summarized in Figure 2.13 (modified after Rohling 1994).

from surface waters (i.e. export production), and (2) reduced deep water ventilation due to diminished excess evaporation from the Mediterranean basin caused by enhanced freshwater discharge (for overviews, see Cita

45°N

40°

A B

35°

C

30°N 5°W

0°

5°E

10°

15°

20°

25°

30°

35°

Present day (pre-1988)

S

N

Adriatic Sea

S

Aegean Sea

N

Depth (m)

0

AeDW

ADW

700

A W

B Open Eastern Mediterranean

E

MIW

AeDW ADW

C Fig. 2.13. Schematic presentation of the changes in subsurface circulation patterns between the present day and times of sapropel formation. The three profiles presented summarize information obtained from analytical and modelling studies from north to south through the Adriatic and Aegean basins, and from west (Strait of Sicily) to east (near Cyprus) through the open eastern Mediterranean. MIW stands for Mediterranean Intermediate Water; ADW for Adriatic Deep Water; AeDW for Aegean Deep Water; AIW for Adriatic Intermediate Water; AeIW for Aegean Intermediate Water; ODW for Old (isolated) Deep Water. Modified from Rohling (1994) and Myers et al. (1998).

The Marine Environment: Present and Past

S

Adriatic Sea

Sapropel-time S N

Aegean Sea

53

N

0

Depth (m)

AeIW AIW

ODW

700

ODW

B

A W

E

Open Eastern Mediterranean AeIW AIW

ODW C Fig. 2.13. (cont.)

Evidence for elevated export production has been compiled through a combination of proxy data that includes phyto- and zoo-plankton abundances, stable isotope gradients, organic-carbon accumulation and composition, and Ba/Al ratios in the sediment (among many others: Cita and Grignani 1982; Rohling and Gieskes 1989; Castradori 1993; Higgs et al. 1994; Thomson et al. 1995, 1999; Van Os et al. 1994; Kemp et al. 1999; Mercone et al. 2000, 2001; Rohling et al. 2004). Productivity increases during the deposition of sapropels appear to have been of a (temporally integrated) basin-wide nature, especially in the form of a Deep Chlorophyll Maximum (Rohling and Gieskes 1989; Castradori 1993; Rohling 1994; Kemp et al. 1999; Corselli et al. 2002), although on shorter time-scales there may have been considerable spatial ‘patchiness’ (Casford et al. 2003). The development of a Deep Chlorophyll Maximum with high export production at sapropel times has been ascribed to hydrographic rearrangements in response to the decrease in buoyancy loss from the Mediterranean at times of enhanced freshwater input. Notably, the reduced surface buoyancy loss is thought to have caused a shoaling of the surface–intermediate water interface from its present

depth below the zone of light penetration (euphotic zone) to a depth within the euphotic layer (Rohling and Gieskes 1989; Rohling 1991b, 1994; Myers et al. 1998). This would allow nutrients stored within the subsurface waters to become utilized for production at the base of the euphotic layer. There is a continuing debate about the ultimate supply of the nutrients that could have supported extensive organic carbon burial in the sediments. Early work concentrated on riverine nutrient input at times of sapropel deposition, but biogeochemical modelling suggests that river-input would be insufficient if the nutrient budget were at steady state during sapropel formation (Stratford et al. 2000). However, recent work has suggested that the basin may have accumulated nutrients over as much as 1,500 years prior to the onset of organic carbon burial, so that the nutrient budget during sapropel deposition ought to be considered as a product of accumulation over much longer timescales, and so was not at steady state (Casford et al. 2002). Strong evidence for enhanced freshwater influx into the eastern Mediterranean at sapropel times comes from negative anomalies in stable oxygen isotope ratios measured on the calcium-carbonate shells of

54

Eelco Rohling et al. Astronomical cycles

+2.0

0

–2.0

Sapropel geochronology

Oxygen Isotope Record G. ruber

Age (Ma)

Sapropels

Depth (m)

Core RC9-181

Precession

–

Eccentricity

+ min

max

S1 1 2 3 4 5 6

S2 S3 S4 S5

0.1

S6 S7 S8 S9

0.2

0.3

7 S10 8

0.4 S11

9 S12

planktonic foraminifera that live in near-surface habitats. Fresh water has distinctly low oxygen isotope ratios compared with sea water, and especially the fresh waters derived from heavy (monsoon-type) rainfalls that are isotopically very light. Freshwater floods to the Mediterranean surface waters therefore cause light isotope anomalies in the surface-dwelling foraminifera (e.g. Vergnaud-Grazzini et al. 1977; Thunell and Williams 1983, 1989; Jenkins and Williams 1984; Ganssen and Troelstra 1987; Kallel et al. 1997a, b; Tang and Stott 1993; Rohling and De Rijk 1999a, b; Emeis et al. 1998, 2000, 2003; Rohling et al. 2004). Sedimentary Ti/Al ratios, palaeomagnetic data, and clay mineralogical studies confirm that times of sapropel deposition were characterized by humid climates with high runoff, whereas intervening times were arid with reduced riverine and enhanced wind-blown sediment supply (e.g. Krom et al. 1999; Foucault and Mélières 2000; Wehausen and Brumsack 2000; Lourens et al. 2001; Larrasoaña et al. 2003). A particularly important discovery concerned the temporal coincidence between sapropel occurrences and insolation-driven monsoon maxima, affecting the eastern Mediterranean via changes in Nile discharge (Rossignol-Strick et al. 1982; Rossignol-Strick 1983, 1985). These authors approached the problem by specifying an index for monsoon intensity (‘monsoon index, M’) as a function of two parameters, namely insolation

Fig. 2.14. Phase relationships between the sapropel record and associated ‰18 O record from core RC9-181 and the orbital cycles of precession and eccentricity (modified from Hilgen 1991a).

at the (north) Tropic of Cancer (IT ), and the insolation difference between the Tropic of Cancer and the equator (IT –IE ), so that M = 2IT –IE . The variation in the index value was considered relative to the value of AD 1950 (Rossignol-Strick 1985). This pioneering work instigated an intensive search into the timing of sapropel formation over their full temporal range, which confirmed that sapropels were always formed at times when perihelion falls in boreal summer (‘precession minima’, relative to ‘maxima’ that represent the present configuration with perihelion in boreal winter). It was also observed that not all precession minima have sapropels, but that they instead occur in discrete clusters. Each cluster was found to represent times of maximum orbital eccentricity, in agreement with eccentricity modulation of the impact of precession (Hilgen 1991a , b; Hilgen et al. 1993, 1995; Lourens et al. 1996, 2001) (Figure 2.14). Numerical climate modelling corroborated the impact of precession and eccentricity on monsoon intensity (e.g. Kutzbach 1985; Kutzbach and Guetter 1986; COHMAP 1988). In essence, insolation changes affect the monsoons: air rises over a hot surface, giving low surface pressure, while it descends over a cool surface, giving high surface pressure. Because land has much lower thermal inertia than ocean, land surfaces experience a much stronger annual fluctuation in both temperature and pressure

The Marine Environment: Present and Past

than ocean surfaces. During periods with enhanced seasonal insolation contrasts, the higher summer insolation increases surface temperatures especially over land, and this in turn amplifies the atmospheric pressure differences between land and sea. In addition to this direct radiative forcing, the preceding winter conditions also play a role, due to the thermal inertia of the ocean. The slow response of oceanic temperatures on seasonal timescales amplifies the land–sea temperature contrast from direct solar heating, and thus enhances the land– sea pressure differences. In summer, the strong land (low) to sea (higher) pressure difference leads to surface air flow from ocean to land. This air flow is moistureladen, because of evaporation over the ocean. The air expands and cools as it rises over the land, a process that is accelerated if the air masses are forced up by mountain ranges. The cooling causes the air masses to shed their vapour as rain. Condensation releases heat, which amplifies the process by enhancing the ascending motion in the air column. Thus, a zone develops of highfrequency and high-intensity monsoonal rains. We emphasize that the above description of the summer monsoon in terms of surface thermal forcing (i.e. as a super sea breeze) represents a strongly simplified generalization. In reality, the low-pressure cell over land derives much of its intensity and continuity from dynamical effects related to the mean high-level wind flow in the atmosphere (at the 500 millibar level, or approximately at 5.5 km height), especially in the case of the Indian/Asian monsoon. As an extra complication, it is thought that the strength of the trade winds in the opposite (winter) hemisphere may determine a ‘push’ across the equator into the summer monsoonal low. Despite its schematic nature, however, the thermal concept offers a reasonable representation of the general features of the African monsoon. Over Africa, the axis of low pressure at the surface (‘the monsoonal low-pressure trough’) follows the seasonal march of the sun at its high point (zenith), which reaches the Tropic of Cancer at the summer solstice. This seasonal swing over the band of monsoon-influenced latitudes in Africa can be smooth because most of Sahelian and Saharan North Africa is relatively flat. The influence of ‘push’ effects by the Southern (winter) Hemisphere trade winds on the North African summer monsoon was accounted for in the monsoon intensity calculations of Rossignol-Strick and co-workers by inclusion of an austral winter insolation gradient (GS ) between 20 and 70◦ S, so that GS = I20 –I70 (Rossignol-Strick 1985). Within the context of monsoon intensification, it is relevant to briefly review reconstructions of the Nile and the Sahara since the LGM, as summarized by Adamson

55

et al. (1980) and Williams et al. (2000). These authors report that, from the LGM until roughly 12,500 years BP , Nile discharge was very low. The White Nile was a seasonal, intermittent river until ∼12,500 years BP, when Lake Victoria overflow developed and the ‘buffering’ Sudd swamps in Sudan became established, which ensured a more regular, perennial discharge from the White Nile. From ∼12,500 years BP, there was an (intermittent) period with very high discharge, associated with the early–mid Holocene monsoon maximum that developed during the insolation maximum of that time. This maximum ended with a development towards generally much drier conditions around 5,000 years BP, heralding the development of the modern Nile regime. These trends are supported by general findings that astronomical forcing affects not only the intensity of the African monsoon, but also its spatial influence, causing strong reductions in the size of the Sahara desert by northward migration of its southern margin. The ‘greening’ of the Sahara is a well-known response in numerical climate models that include vegetation– climate feedback mechanisms (Brovkin et al. 1998; Claussen et al. 1998). The concept is supported by a wide variety of field observations: rock-art and animal, human, and vegetation remains in the central Sahara; a massive expansion of Lake Chad; the presence of substantial palaeolakes in currently hyperarid areas such as the Oyo depression of NW Sudan; and activation of large-scale systems of presently inactive wadis (e.g. Pachur and Braun 1980; Gaven et al. 1981; Ritchie et al. 1985; McKenzie 1993; Szabo et al. 1995; PetitMaire and Guo 1997; Pachur 2001; Gasse 2000, Hoelzmann et al. 2000; Williams et al. 2000; Mandel and Simmons 2001; Hassan 2002; and many references therein). A recent study suggested on the basis of stable oxygen isotope data that the monsoon front penetrated sufficiently far northwards during the insolation maximum of the previous interglacial maximum to have caused significant precipitation to the north of the central Saharan watershed (∼21 ◦ N) (Rohling et al. 2002b). In that case, significant runoff would have reached not only the eastern Mediterranean via the Nile River, but also the wider North African margin. A study concerned with aeolian dust variations over the last 3 million years supported that scenario for all substantial insolation maxima (Larrasoaña et al. 2003). Archaeological observations around exclusively rain-fed depressions on the Libyan Plateau suggest that monsoonal summer rains of central Africa periodically penetrated at least as far north as Kharga (roughly 25 ◦ N) during the early–mid Holocene, despite the fact that conditions during that pluvial phase seem to have remained

56

Eelco Rohling et al.

drier than during earlier Quaternary pluvial phases (Mandel and Simmons 2001). The observation of Mandel and Simmons (2001) that the Holocene monsoon maximum was of a relatively low intensity compared with previous Quaternary monsoon maxima has been corroborated by recent work to quantify the Holocene and previous interglacial (Eemian) monsoon impacts on the freshwater budget of the eastern Mediterranean (Rohling 1999b; Rohling et al. 2004). Eastern Mediterranean surface-water oxygen isotope (‰18 O) data show two very distinct ‘peaks’ in the Eemian monsoon maximum, separated by an ‘interruption’ that lasted about 800 years. A mixed-layer ‰18 O box model to quantify freshwater flooding during the two Eemian monsoon peaks suggests that the basinaveraged Mediterranean excess of evaporation over freshwater input was reduced to 5–45 per cent (older peak) and 35–60 per cent (younger peak), relative to the present (Rohling et al. 2004). It also suggests that the interruption between the two peaks was characterized by excess evaporation at levels very close to present-day values. Using a similar technique, the excess evaporation during the early–mid Holocene monsoon maximum was estimated at about 65 per cent of the present value (Rohling 1999b). Using a different technique, based on an ocean general circulation model (OGCM), Myers (2002) suggests that the excess evaporation value for the Holocene monsoon maximum ranged below 80 per cent and most likely around 20–40 per cent of the present-day value. As yet, eastern Mediterranean ‰18 O records through the Holocene monsoon maximum have revealed only weak indications of a monsoon interruption, but neodymium (Nd) isotopes seem more conclusive that such an interruption did occur during the Holocene (Scrivner et al. 2004). African lake levels also clearly demonstrate an arid interlude, dated between about 8.5 and 7.8 ka BP (Gasse 2000), coincident with a cooling event of ∼500-year duration over the Aegean and Adriatic seas (see e.g. Rohling et al. 2002a ). There is a growing body of evidence showing that it was not just the monsoon system that was intensified at times of sapropel deposition. Records of pollen and spore abundances from terrestrial vegetation suggest high abundances of species that require wet summers around the Northern Borderlands of the Eastern Mediterranean (NBEM) at times of sapropel formation (Rossignol-Strick 1987, 1995; Wijmstra et al. 1990; Rohling and Hilgen 1991; Tzedakis 1993; Mommersteeg et al. 1995; Frogley et al. 1999). Winter precipitation is thought to have been increased in the NBEM as well (Wijmstra et al. 1990). Isotope studies on speleothems corroborate

the inferred increase in precipitation (Bard et al. 2002; Matthews et al. 2000; Bar-Matthews et al. 1999, 2000, 2003), as do elevated lake levels (e.g. Digerfeldt et al. 2000). The conditions at times of sapropel formation would therefore appear to have been considerably different from the typical dry-summer climate that characterizes the area today. Direct precipitation from the African and Indian monsoon systems is unlikely to have penetrated into the Mediterranean basin, demonstrating that the moisture for precipitation in the NBEM derived from regional processes, probably in the form of Mediterranean depressions (Rohling and Hilgen 1991). This notion was corroborated by isotopic characteristics of speleothems in Soreq Cave, Israel, which demonstrate a local Mediterranean moisture origin (Matthews et al. 2000; Bar-Matthews et al. 2003). The summer flux of Mediterranean moisture at times of insolation maxima affected even the northernmost tip of the Red Sea, where it has been described as the ‘Mediterranean monsoon’ (Arz et al. 2003). Importantly, this process would not likely have affected the Mediterranean Sea’s overall hydrological budget very significantly—although any precipitation would have led to runoff into the basin, the original evaporative loss took place from the same basin. Any transport into another basin’s watershed area (e.g. the Red Sea, Jordan Valley and Dead Sea, or Tigris/Euphrates and Persian Gulf) would imply that the regional ‘humidity’ in the NBEM might even have coincided with a slight increase in net evaporative loss from the Mediterranean. Importantly, however, the process would reflect considerable freshwater redistribution in a generally eastward direction within the Mediterranean basin, so that the hydrological budget may have been substantially affected on local scales and in terms of regional gradients. The (especially monsoon-related) reduction of Mediterranean excess evaporation would have caused a reduction in the salinity of newly formed intermediate water. Numerical modelling suggests that intermediatewater formation is likely to have shifted from a normal salt-dominated LIW mode, to a temperature-dominated mode driven from the Adriatic Sea at times of sapropel formation (Myers et al. 1998; Myers 2002). Stable isotope data for planktonic foraminiferal species with different depth habitats from the Aegean Sea corroborate that notion (Casford et al. 2002). The collapse of the first, salt-driven, stage of the deep-ventilation ‘motor’ would have caused any new deep water to form at much lower salinities (hence, lower densities) than it did in times before the monsoon intensification (Rohling 1994; Myers et al. 1998; Myers 2002). Thus,

The Marine Environment: Present and Past

newly formed deep water masses could not displace the existing, denser, ‘old’ deep waters (ODW) formed before the freshwater flooding (Figure 2.13). As a result, the ODW became poorly ventilated, and eventually oxygen depleted due to continuing remineralization of sinking organic matter (for an overview, see Rohling 1994). At least down to ∼2000 m depth some occasional ventilation may have persisted during the deposition of several sapropels, and down to those depths the occurrence of truly anoxic conditions was probably restricted to spatially discontinuous ‘blankets’ over the sea-floor topography (Casford et al. 2003).

Centennial- to Millennial-scale Variability As mentioned previously, deep ventilation in the Mediterranean basin is strongly affected by wintertime, orographically channelled, northerly outbursts of cold polar and continental air over the northern sectors of the basin. Fluctuations in the intensity and frequency of such events are also reflected in temperature proxy data. Terrestrial and marine Mediterranean palaeoclimate and palaeoceanographic proxy records that are resolved on centennial timescales have been found to reflect multi-centennial to millennial fluctuations (Rohling et al. 1998b, 2002a, b, 2004; Paterne et al. 1999; Cacho et al. 1999, 2000, 2001, 2002; Allen et al. 1999; Combourieu-Nebout et al. 2002; SánchezGoñi et al. 2002; Tzedakis et al. 2004). These have often been related to climatic oscillations in the wider North Atlantic realm by correlation with the ice-‰18 O records and non-sea-salt ion series (dust) from the welldated GISP2 and GRIP ice cores (Greenland summit). Particularly strong cooling events have been described for the northern sectors of the Mediterranean at times coincident with the North Atlantic ‘Heinrich Events’ of massive ice-rafting (Rohling et al. 1998b; Paterne et al. 1999; Cacho et al. 1999, 2000, 2001, 2002). Within the last glacial cycle, periods of intensified or more frequent northerly cooling events in the Mediterranean generally correlate with the DansgaardOeschger (DO) stadials (cold episodes), and the most intense events correlate with the most intense DO stadials, which were marked in the North Atlantic by the Heinrich Events. Highly resolved records from the Mediterranean region furthermore indicate that the cool periods were characterized by enhanced aridity (e.g. Allen et al. 1999; Sánchez-Goñi et al. 2002; Combourieu-Nebout et al. 2002; Tzedakis et al. 2004; Hoogakker et al. 2004).

57

The most recent sapropel (S1) formed between about 9,000 and 6,000 calibrated years BP, in association with the monsoon maximum of the current interglacial period. Detailed work has found an ‘interruption’ in the deposition of the S1 sapropel, marking a period of improved deep water ventilation that spans several centuries around 8,500–8,000 years BP (van Straaten 1966, 1970, 1972, 1985; Rohling et al. 1997; De Rijk et al. 1999; Geraga et al. 2000; Mercone et al. 2000, 2001; Casford et al. 2001, 2002, 2003; Rohling et al. 2002a ). This ventilation event coincided with intensified cooling over the Adriatic and Aegean seas, which has been related to an increase in the intensity or frequency of northerly cold outbursts over those regions, in association with a widespread North Atlantic cold event (Rohling et al. 2002a ). Similarly, the ending of sapropel deposition was marked by a cooling, around 6,000 years BP (Rohling et al. 1997; Geraga et al. 2000; Casford et al. 2001, 2002, 2003; Mercone et al. 2001; Rohling et al. 2002a). Pollen data confirm that the cooling periods were normally marked by enhanced aridity (e.g. Rossignol-Strick 1995; Geraga et al. 2000). Episodically improved deep water ventilation within times of generally poor ventilation (sapropel conditions) has since been inferred for a large number of sapropels, suggesting that climatic variability on short timescales persisted even in these periods of generally warm and humid conditions in the Mediterranean region (Casford et al. 2003). The above might give the impression that sapropels resulted from monsoon maxima extending over several millennia, while some internal variability occurred due to intermittent cooling events originating from the north, in association with polar or temperate climate events. This would be misleading, because the centennial-scale episodes of increased cooling from the north are well known to have been associated with severe reductions in monsoon runoff. One likely example is the 800-year monsoon interruption within Eemian sapropel S5 (e.g. Cane et al. 2002; Rohling et al. 2002b, 2004). Monsoon-fed African lakes show a similar interruption within the Holocene monsoon maximum, as part of a series of distinct and abrupt periods of low lake levels that coincide in time with the northerly coolings recorded in the Mediterranean, as described above. The particularly pronounced period of low lake levels between 8.5 and 7.8 ka BP (Gasse 2000) coincides closely with the interruption of sapropel S1, and recent compilations have found this to be a widespread interval of climate deterioration throughout (at least) the Northern Hemisphere (Mayewski et al. 2004; Rohling and Pälike 2005). Furthermore, Egyptian archaeological

58

Eelco Rohling et al.

records for the Holocene indicate dramatic turnovers related to strong fluctuations in intensity and frequency of Nile flooding (Hassan 1997, 2002, and references therein). Clearly, some fundamental, if elusive, connection exists between cooling from the north and reductions in (African) monsoon intensity. Hence, it is worthwhile to spend some time looking at variability in the wind-blown dust flux into the Mediterranean, as a measure of North African climate variability. At present, the northward transport of aeolian (windblown) dust over the Mediterranean is linked to the presence of cyclones over the basin (Moulin et al. 1997), and most Saharan dust deposition over southern Europe occurs with precipitation (Bücher and Lucas 1984; Bergametti et al. 1989a; Loye-Pilot et al. 1989; Guerzoni et al. 1992). Important source areas of dust transport to Western Europe are Algeria, the Western Sahara, and the Moroccan Atlas (Molinaroli 1996; Avila et al. 1997; Goudie and Middleton 2001; Chapter 14). Weldeab et al. (2002) use Si/Al and Ti/Al ratios as well as Sr and Nd isotopes to show that the Saharan terrigenous input into the western Mediterranean Sea is predominantly from the southwest (Morocco/north-western Algeria) and south-east (Tunisia/western Libya) during interglacial periods and from the southern Saharan/Sahelian region during glacial times. Eastern Libya and Egypt are the important source areas of aeolian dust to the eastern Mediterranean basin. Overall, terrigenous input into the Mediterranean at glacial times greatly exceeded that of interglacial times (Weldeab et al. 2002). Fluvial sediment yields were also higher (Macklin et al. 2002). A continuous 3 million-year record of dust supply from the northern Sahara into the eastern Mediterranean, developed from sediment magnetic data, consistently shows dust-flux minima at times of Northern Hemisphere insolation/monsoon maxima (Larrasoaña et al. 2003). These minima were related to northward penetration of the African summer monsoon front beyond the central Saharan watershed (∼21 ◦ N), as proposed previously on the basis of Mediterranean oxygen isotope data (Rohling et al. 2002b). Such northward penetration of the African summer monsoon agrees with a broad expansion of (savannah-like) vegetation cover shown in both observations and modelling experiments (‘greening of the Sahara’: Claussen et al. 1998; Brovkin et al. 1998; Irizarry-Ortiz et al. 2003; and references therein). Consequently increased soil cohesiveness throughout large areas of the northern Sahara would cause a decrease in dust production, similar to modern conditions in the Sahel (Middleton 1985).

A high-resolution record of lithogenic fraction variability from the Alboran Sea has revealed millennialto submillennial-scale oscillations. These correlate with Atlantic Dansgaard-Oeschger stadials and Heinrich Events, and were characterized by increases in the northward transport of Saharan dust (Moreno et al. 2002). Similar increases in Saharan dust supply at times of DO stadials have been inferred from sediment cores throughout the wider western Mediterranean (Hoogakker et al. 2004), and detailed magnetic susceptibility records for cores taken in the eastern Mediterranean (pp. 3/28–3/31 in Hemleben et al. 2003) suggest that a millennial-scale aeolian dust signal may also be preserved in that basin. At this stage, we do not infer that the monsoon penetration model inferred for the longer-term (Milankovitch-scale) dust cycles should apply to the shorter-term (sub-Milankovitch) events. Instead, the shorter-term dust-flux variations may well be controlled by cyclone activity, as it appears to be on interannual to decadal time scales (Moulin et al. 1997). An important control on cyclogenesis within the Mediterranean basin is exerted by cold (arid) air outbursts over the northern sectors (Chapter 3), and this may be the mechanism underlying the correlation between dust cycles in the Mediterranean and DO events in the North Atlantic region. This interpretation remains speculative, however, until more process-oriented research leads to a more detailed understanding.

References Adamson, D. A., Gasse, F., Street, F. A., and Williams, M. A. J. (1980), Late Quaternary history of the Nile. Nature 288: 50–5. Allen, J. R., Brandt, U., Brauer, A., Hubberten, H. W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J. F. W., Nowaczyk, N. R., Oberhansli, H., Watts, W. A., Wulf, S., and Zolitschka, B. (1999), Rapid environmental changes in southern Europe during the last glacial period. Nature 400: 740–3. Arnone, R. A., Wiesenburg, D. A., and Saunders, K. (1990), The origin and characteristics of the Algerian current. Journal of Geophysical Research 95: 1587–98. Arz, H. W., Lamy, F., Pätzold, J., Müller, P. J., and Prins, M. (2003), Mediterranean moisture source for an earlyHolocene humid period in the Northern Red Sea. Science 300: 118–21. Avila, A., Queralt-Mitjans, L., and Alarcón, M. (1997), Mineralogical composition of African dust delivered by red rains over northeastern Spain. Journal of Geophysical Research 102: 21977–96. Bard, E., Delaygue, G., Rostek, F., Antonioli, F., Silenzi, S., and Schrag, D. P. (2002), Hydrological conditions over the western Mediterranean basin during the deposition of the cold Sapropel 6 (ca. 175 kyr BP). Earth and Planetary Science Letters 202: 481–94.

The Marine Environment: Present and Past Bar-Matthews, M., Ayalon, A., Kaufman, A., and Wasserburg, G. J. (1999), The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166: 85–95. (2000), Timing and hydrological conditions of Sapropel events in the Eastern Mediterranean, as evident from speleothems, Soreq cave, Israel. Chemical Geology 169: 145–56. Gilmour, M., Matthews, A., and Hawksesworth, C. (2003), Sea–land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implication for paleorainfall during interglacial intervals. Geochimica et Cosmochimica Acta 67: 3181–99. Bergametti, G., Dutot, L., Buat-Ménard, P., and Remoudaki, E. (1989), Seasonal variability of the elemental composition of atmospheric aerosol particles over the northwestern Mediterranean. Tellus 41B: 553–61. Béthoux, J. P. (1984), Paleo-hydrologie de la Mer Mediterranee au cours des derniers 20 000 ans. Oceanologica Acta 7: 43–8. and Gentili, B. (1994), The Mediterranean Sea, a test area for marine and climatic interactions, in P. Malanotte-Rizzoli and A. R. Robinson (eds.), Ocean Processes in Climate Dynamics: Global and Mediterranean Examples. Kluwer, Dordrecht, 239–54. Morin, P., Nicolas, E., Pierre, C., and Ruiz-Pino, D. (1999) The Mediterranean Sea: a miniature ocean for climatic and environmental studies and a key for the climatic functioning of the North Atlantic. Progress in Oceanography 44: 131–46. Boucher, K. (1975) Global Climate. English University Press, London. Broecker, W. S. (2000), Abrupt climate change: causal constraints provided by the paleoclimate record. Earth Science Reviews 51: 137–54. Brovkin, V., Claussen, M., Petoukhov, V., and Ganopolski, A. (1998), On the stability of the atmosphere–vegetation system in the Sahara/Sahel region. Journal of Geophysical Research 103: 31613–24. Bruce, J. G. and Charnock, H. (1965), Studies of winter sinking of cold water in the Aegean Sea. Rapports Com. Internat. Mer Méditerranée 18: 773–8. Bryden, H. L. and Kinder, T. H. (1991), Steady two-layer exchange through the Strait of Gibraltar. Deep-Sea Research 38: S445–S463. Candela, J., and Kinder, T. H. (1994), Exchange through the Strait of Gibraltar. Progress in Oceanography 33: 201–48. Bücher, A. and Lucas, G. (1984), Sédimentation éolienne intercontinentale, poussières sahariennes et géologie. Bull. Centr. Rech. Expl.-Prod. Elf-Aquit. 8: 151–65. Burman, I. and Oren, O. H. (1970), Water outflow close to the bottom from the Aegean. Cahiers Océanographique 22: 775–80. Cacho, I., Grimalt, J. O., Pelejero, C., Canals, M., Sierro, F. J., Flores, J. A., and Shackleton, N. J. (1999), Dansgaard-Oeschger and Heinrich event imprints in Alboran Sea paleotemperatures. Paleoceanography 14: 698–705. Sierro, F. J., Shackleton, N. J., and Canals, M. (2000), Evidence of enhanced Mediterranean thermohaline circulation during rapid climate coolings. Earth and Planetary Science Letters 183: 417–29.

59

Canals, M., Sbaffi, L., Shackleton, N. J., Schönfeld, J., and Zahn, R. (2001), Variability of the western Mediterranean Sea surface temperatures during the last 25,000 years and its connection with the northern hemisphere climatic changes. Paleoceanography 16: 40–52. and Canals, M. (2002), Response of the Western Mediterranean Sea to rapid climatic variability during the last 50,000 years: a molecular biomarker approach. Journal of Marine Systems 33–4: 253–72. Cane, T., Rohling, E. J., Kemp, A. E. S., Cooke, S., and Pearce, R. B. (2002), High-resolution stratigraphic framework for Mediterranean sapropel S5: defining temporal relationships between records of Eemian climate variability. Palaeogeography, Palaeoclimatology, Palaeoecology 183: 87–101. Cantu, V. (1977), The climate of Italy, in C. C. Wallen (ed.), Climates of Central and Southern Europe, World Survey of Climatology 6. Elsevier, Amsterdam, 127–83. Casford, J. S. L., Abu-Zied, R. H., Rohling, E. J., Cooke, S., Boessenkool, K. P., Brinkhuis, H., de Vries, C., Wefer, G., Geraga, M., Papatheodorou, G., Croudace, I., Thomson, J., Wells, N. C., and Lykousis, V. (2001), Mediterranean climate variability during the Holocene. Mediterranean Marine Science 2: 45–55. Rohling, E. J., Abu-Zied, R. H., Cooke, S., Fontanier, C., Leng, M., and Lykousis,V. (2002), Circulation changes and nutrient concentrations in the Late Quaternary Aegean Sea: A nonsteady state concept for sapropel formation. Paleoceanography 17, doi 10.1029/2000PA000601. Jorissen, F. J., Leng, M., and Thomson, J. (2003), A dynamic concept for eastern Mediterranean circulation and oxygenation during sapropel formation. Palaeogeography, Palaeoclimatology, Palaeoecology 190: 103–19. Castradori, D. (1993), Calcareous nannofossils and the origin of eastern Mediterranean sapropels. Paleoceanography 8: 459–71. Cheney, R. E. and Doblar, R. A. (1982), Structure and variability of the Alboran Sea frontal system. Journal of Geophysical Research 87: 585–94. Cita, M. B. (1973), Inventory of biostratigraphical findings and problems, in W. B. F. Ryan, K. J. Hsü, et al. (eds.). Initial Reports of the Deep-Sea Drilling Project 13. US Govt. Printing Office, Washington, DC), 1045–73. and Grignani, D. (1982), Nature and Origin of Late Neogene Mediterranean Sapropels, in S. O. Schlanger and M. B. Cita (eds.) Nature and Origin of Cretaceous Carbon-Rich Facies. Academic Press, San Diego, 165–96. Vergnaud-Grazzini, C., Robert, C., Chamley, H., Ciaranfi, N., and d’Onofrio, S. (1977), Paleoclimatic record of a long deep sea core from the eastern Mediterranean. Quaternary Research 8: 205–35. Claussen, M., Brovkin, V., Ganopolski, A., Kubatski, C., and Petoukho, V. (1998), Modelling global terrestrial vegetationclimate interaction. Philosophical Transactions Royal Society, London 353: 53–63. Clemens, S. C. and Prell, W. L. (1990), Late Pleistocene variability of Arabian Sea summer monsoon winds and continental aridity: eolian records from the lithogenic component of deep-sea sediments. Paleoceanography 5: 109–45. Murray, D. W., and Prell, W. L. (1996), Nonstationary phase of the Plio-Pleistocene Asian monsoon. Science 274: 943–8.

60

Eelco Rohling et al.

COHMAP members (1988), Climatic changes of the last 18,000 years: observations and model simulations. Science 241: 1043–52. Combourieu-Nebout, N., Turon, J. L., Zahn, R., Capotondi, L., Londeix, L., and Pahnke, K. (2002), Enhanced aridity and atmospheric high-pressure stability over the western Mediterranean during the North Atlantic cold events of the past 50 k.y. Geology 30: 863–6. Corselli, C., Principato, M. S., Maffioli, P., and Crudelli, D. (2002), Changes in planktonic assemblages during sapropel S5 deposition: Evidence from Urania Basin area, eastern Mediterranean. Paleoceanography 17: doi 10.1029/2000PA000536. Cramp, A. and O’Sullivan, G. (1999), Neogene sapropels in the Mediterranean: a review. Marine Geology 153: 11–28. Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjornsdottir, A. E., Jouzel, J., and Bond, G. (1993), Evidence for general instability of past climate from a 250 kyr ice core. Nature 364: 218–19. De Rijk, S., Rohling, E. J., and Hayes, A. (1999), Onset of climatic deterioration in the eastern Mediterranean around 7 ky BP; micropalaeontological data from Mediterranean sapropel interruptions. Marine Geology 153: 337–43. Digerfeldt, G., Olsson, S., and Sandgren, P. (2000), Reconstruction of lake-level changes in lake Xinias, central Greece, during the last 40,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 65–82. Dinarès-Turell, J., Hoogakker, B. A. A., Roberts, A. P., Rohling, E. J., and Sagnotti, L. (2003), Quaternary climatic control of biogenic magnetite production and eolian dust input in cores from the Mediterranean Sea. Palaeogeography, Palaeoclimatology, Palaeoecology 190: 195–209. Dünkeloh, A. and Jacobeit, J. (2003), Circulation dynamics of Mediterranean precipitation variability 1948–1998, International Journal of Climatology 23: 1843–66. El Dardir, M. (1994), Sedimentation in the Nile High Dam reservoir, 1987–1992, and sedimentary futurologic aspects. Sedimentology of Egypt 2: 23–39. Elenga, H., Peyron, O., Bonnefille, R., Jolly, D., Cheddadi, R., Guiot, J., Andrieu, V., Bottema, S., Bucher, G., de Beaulieu, J. L., Hamilton, A. C., Maley, J., Marchant, R., Perez-Obiol, R., Reille, M., Riollet, G., Scott, L., Straka, H., Taylor, D., Van Campo, E., Vincens, A., Laarif, F., and Jonson, H. (2000), Pollen-based biome reconstruction for southern Europe and Africa, 18,000 yr BP. Journal of Biogeography 27: 621–34. Emeis, K. C., Camerlenghi, A., McKenzie, J. A., Rio, D., and Sprovieri, R. (1991), The occurrence and significance of Pleistocene and Upper Pliocene sapropels in the Tyrrhenian Sea, Marine Geology 100: 155–82. Sakamoto, T., Doose, H., Erlenkeuser, H., Howell, M., Kroon, D., and Paterne, M. (1998), Stable isotope and temperature records of sapropels from ODP Sites 964 and 967: Constraining the physical environment of sapropel formation in the Eastern Mediterranean Sea, in A. H. F. Robertson, K. C. Emeis, C. Richter, and A. Camerlenghi (eds.), Proceedings ODP, Scientific Results 160. Ocean Drilling Program, College Station, Tex., 309–31. Struck, U., Schulz, H. M., Bernasconi, S., Sakamoto, T., and Martinez-Ruiz, F. (2000), Temperature and salinity of Mediterranean Sea surface waters over the last 16,000 years: Constraints on the physical environment of S1 sapropel formation based on stable oxygen isotopes and alkenone unsaturation

ratios. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 259–80. Schulz, H., Struck, U., Rossignol-Strick, M., Erlenkeuser, H., Howell, M. W., Kroon, D., Mackensen, H., Ishizuka, S., Oba, T., Sakamoto, T., and Koizumi, I. (2003), Eastern Mediterranean surface water temperatures and ‰18 O composition during deposition of sapropels in the late Quaternary. Paleoceanography 18: doi 1029/2000PA000617. Fairbanks, R. G. (1989), A 17,000 year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation, Nature 342: 637–42. (1990), The age and origin of the ‘Younger Dryas Climate Event’ in Greenland Ice Cores. Paleoceanography 5: 937–48. Foucault, A. and Mélières, F. (2000), Palaeoclimatic cyclicity in central Mediterranean Pliocene sediments: the mineralogical signal. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 311–23. Frogley, M. R., Tzedakis, P. C., and Heaton, T. H. E. (1999), Climate variability in Northwest Greece during the last interglacial. Science 285: 1886–9. Ganssen, G. and Troelstra, S. R. (1987), Palaeoenvironmental changes from stable isotopes in planktonic foraminifera from Eastern Mediterranean Sapropels. Marine Geology 75: 221–30. Garrett, C. (1994), The Mediterranean Sea as a climate test basin, in P. Malanotte-Rizzoli and A. R. Robinson (eds.), Ocean Processes in Climate Dynamics: Global and Mediterranean Examples. Kluwer, Dordrecht, 227–37. Garzoli, S. and Maillard, C. (1979), Winter circulation in the Sicily and Sardinia straits region. Deep-Sea Research 26A: 933–54. Gascard, J. C. (1978), Mediterranean deep water formation, baroclinic instability and oceanic eddies. Oceanologica Acta 1: 315–30. and Richez, C. (1985), Water mass and circulation in the western Alboran Sea and in the Strait of Gibraltar. Progress in Oceanography 15: 157–216. Gasse, F. (2000), Hydrological changes in the African tropics since the last glacial maximum, Quaternary Science Reviews 19: 189–211. Gaven, C., Hillaire-Marcel, C., and Petit-Maire, N. (1981), A Pleistocene lacustrine episode in southeastern Libya. Nature 290: 131–3. Geraga, M., Tsaila-Monopolis, S. Ioakim, Chr., Papatheodorou, G., and Ferentinos, G. (2000), Evaluation of palaeoenvironmental changes during the last 18,000 years in the Myrtoon basin, SW Aegean Sea. Paleogeography, Palaeoclimatology, Palaeoecology 156: 1–17. Gilman, C. and Garrett, C. (1994), Heat flux parameterizations for the Mediterranean Sea: The role of atmospheric aerosols and constraints from the water budget. Journal of Geophysical Research 99: 5119–34. Goudie, A. S. and Middleton, N. J. (2001), Saharan dust storms: nature and consequences. Earth Science Reviews 56: 179–204. Grootes, P. M., Stuiver, M., White, J. W. C., Johnsen, S., and Jouzel, J. (1993), Comparison of oxygen isotope records from the GISP2 and GRIP Greenland ice cores. Nature 366: 552–4. Guerzoni, S., Landuzzi, W., Lenaz, R., Quarantotto, G., Cesari, G., Rampazzo, R., and Molinaroli, E. (1992), Mineral atmospheric particulate from south to northwest Mediterranean: seasonal variations and characteristics. Water Pollution Research Reports 28: 483–93.

The Marine Environment: Present and Past Hassan, F. A. (1997), The dynamics of a riverine civilization: a geoarchaeological perspective on the Nile Valley, Egypt, World Archaeology 29: 51–74. (ed.) (2002), Droughts, Food, and Culture: Ecological Change and Food Security in Africa’s Later Prehistory. Kluwer/ Plenum, New York. Hayes, A., Kucera, M., Kallel, N., Sbaffi, L., and Rohling, E. J. (2005), Glacial Mediterranean sea surface temperatures reconstructed from planktonic foraminiferal assemblages. Quaternary Science Reviews 24: 999–1016. Heburn, G. W. and La Violette, P. E. (1990), Variations in the structure of the anticyclonic gyres found in the Alboran Sea. Journal of Geophysical Research 92: 2901–6. Hemleben, Ch., Hoernle, K., Jørgensen, B. B., and Roether, W. (2003), Ostatlantik–Mittelmeer–Schwarzes Meer, Cruise 51, 12 September–28 December 2001, Meteor Berichte, 03-1. Hamburg University. Hemming, S. R. (2004), Heinrich events: Massive detritus layers of the North Atlantic and their global climate impact. Reviews in Geophysics 42: RG1005, doi 10.1029/ 2003RG 000128. Higgs, N. C., Thomson, J., Wilson, T. R. S., and Croudace, I. W. (1994), Modification and complete removal of eastern Mediterranean sapropels by postdepositional oxidation. Geology 22: 423–6. Hilgen, F. J. (1991a), Astronomical calibration of Gauss to Matuyama sapropels in the Mediterranean and implication for the geomagnetic Polarity Time Scale. Earth and Planetary Science Letters 104: 226–44. (1991b), Extension of the astronomically calibrated (polarity) timescale to the Miocene/Pliocene boundary. Earth and Planetary Science Letters 107: 349–68. Lourens, L. J., Berger, A., and Loutre, M. F. (1993), Evaluation of the astronomically calibrated time scale for the late Pliocene and earliest Pleistocene. Paleoceanography 8: 549–656. Krijgsman, W., Langereis, C. G., Lourens, L. J., Santarelli, A., and Zachariasse, W. J. (1995), Extending the astronomical (polarity) time scale into the Miocene. Earth and Planetary Science Letters 136: 495–510. Hill, A. E. and Mitchelson-Jacob, E. G. (1993), Observations of a poleward-flowing saline core on the continental slope west of Scotland. Deep-Sea Research 40: 1521–7. Hoelzmann, P., Kruse, H. J., and Rottinger, F. (2000), Precipitation estimates for the eastern Saharan palaeomonsoon based on a water balance model of the West Nubian Palaeolake Basin. Global and Planetary Change 26: 105–20. Hoogakker, B. A. A., Rothwell, R. G., Rohling, E. J., Paterne, M., Stow, D. A. V., Herrle, J. O., and Clayton, T. (2004), Aridity episodes during the last glacial cycle recorded in calcium carbonate records from the western Mediterranean Sea. Marine Geology 211: 21–43. Hughes, P. D., Gibbard, P. L., and Woodward, J. C. (2006), Quaternary glacial history of the Mediterranean mountains. Progress in Physical Geography 30: 334–64. Hurst, H. E. (1944), A Short Account of the Nile Basin. Physical Department, Ministry of Public Works, Egypt, Paper 45. Imbrie, J., Shackleton, N. J., Pisias, N. G., Morley, J. J., Prell, W. L., Martinson, D. G., Hays, J. D., McIntyre, A., and Mix, A. C. (1984), The orbital theory of Pleistocene climate: support from a revised chronology of the marine ‰18 O record, in A. Berger, J.

61

Imbrie, J. Hays, G. Kukla, and B. Saltzman (eds.), Milankovitch and Climate: Understanding the Response to Astronomical Forcing. D. Reidel, Hingham, Mass., 269–305. Boyle, E. A., Clemens, S. C., Duffy, A., Howard, W. R., Kukla, G., Kutzbach, J., Martinson, D. G., McIntyre, A., Mix, A. C., Molfino, B., Morley, J. J., Peterson, L. C., Pisias, N. G., Prell, W. L., Raymo, M. E., Shackleton, N. J., and Toggweiler, J. R. (1992), On the structure and origin of major glaciation cycles. 1. Linear responses to Milankovitch forcing. Paleoceanography 7: 701–38. Iorga, M. C. and Lozier, M. S. (1999), Signatures of the Mediterranean outflow from a North Atlantic climatology 1. Salinity and density fields. Journal of Geophysical Research 104: 25985– 26009. Irizarry-Ortiz, M. M., Wang, G., and Eltahir, E. A. B. (2003), Role of the biosphere in the mid-Holocene climate of West Africa. Journal of Geophysical Research 108: doi 10.1029/2001JD000989. Jenkins, J. A. and Williams, D. F. (1984), Nile water as a cause of eastern Mediterranean sapropel formation: evidence for and against. Marine Micropaleontology 9: 521–34. Jongsma, D., Fortuin, A. R., Huson, W., Troelstra, S. R., Klaver, G. T., Peters, J. M., Van Harten, D., De Lange, G. J., and Ten Haven, L. (1983), Discovery of an anoxic basin within the Strabo Trench, eastern Mediterranean. Nature 305: 795–7. Jorissen, F. J. (1999), Benthic foraminiferal successions across Late Quaternary Mediterranean sapropels. Marine Geology 153: 91–101. Asioli, A., Borsetti, A. M., Capotondi, L., De Visscher, J. P., Hilgen, F. J., Rohling, E. J., Van der Borg, K., VergnaudGrazzini, C., and Zachariasse, W. J. (1993), Late Quaternary central Mediterranean biochronology. Marine Micropaleontology 21: 169–89. Kallel, N., Paterne, M., Duplessy, J. C., Vergnaud-Grazzini, C., Pujol, C., Labeyrie, L., Arnold, M., Fontugne, M., and Pierre, C. (1997a ), Enhanced rainfall in the Mediterranean region during the last sapropel event. Oceanologica Acta20: 697–712. Labeyrie, L., Duplessy, J. C., Arnold, M. (1997b), Temperature and salinity records of the Tyrrhenian Sea during the last 18,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 135: 97–108. Kemp, A. E. S., Pearce, R. B., Koizumi, I., Pike, J., and Rance, S. J. (1999), The role of mat-forming diatoms in the formation of Mediterranean sapropels. Nature 398: 57–61. Klein, B., Roether, W., Manca, B. B., Bregant, D., Beitzel, V., Kovacevic, V., and Luchetta, A. (1999), The large deep water transient in the eastern Mediterranean. Deep-Sea Research 46: 371–414. Krom, M. D., Cliff, R. A., Eijsink, L. M., Herut, B., and Chester, R. (1999), The characterisation of Saharan dust and Nile particulate matter in surface sediments from the Levantine basin using Sr isotopes. Marine Geology 155: 319–30. Kullenberg, B. (1952), On the salinity of the water contained in marine sediments. Medd. Oceanogr. Inst. Göteborg 21: 1–38. Kutzbach, J. E. (1985), Modeling of paleoclimates, in Issues in atmospheric and oceanic modeling, A: climate dynamics, in S. Manabe (ed.), Advances in Geophysics 28: 161–196. Kutzbach, J. E. and Guetter, P. J. (1986), The influence of changing orbital parameters and surface boundary conditions on climate simulations for the past 18,000 years. Journal of Atmospheric Science 43: 1726–59.

62

Eelco Rohling et al.

Langway Jr., C. C., Oeschger, H., and Dansgaard, W. (1985) (eds.), Greenland Ice Core: Geophysics, geochemistry, and the environment. Geophysical Monograph 33. American Geophysical Union, Washington, DC. Larrasoaña, J., Roberts, A. P., Rohling, E. J., Winklhofer, M., and Wehausen, R. (2003), Three million years of monsoon variability over the northern Sahara. Climate Dynamics 21: 689–98. Lascaratos, A. (1989), Hydrology of the Aegean Sea. Reports in Meteorology and Oceanography 40/1: 313–431. Laskar, J. (1999), The limits of Earth orbital calculations for geological time scale use. Philosophical Transactions Royal Society, London, A 357: 1735–59. Joutel, F., and Boudin, F. (1993), Orbital precessional, and insolation quantities for the Earth from –20 to +10 Myr. Astronomy and Astrophysics 270: 522–33. La Violette, P. E. (1986), Short term measurements of surface currents associated with the Alboran Sea during Donde Va? Journal of Physical Oceanography 16: 262–79. Leaman, K. D. and Schott, F. A. (1991), Hydrographic structure of the convection regime in the Gulf of Lions: Winter 1987. Journal of Physical Oceanography 21: 575–98. Lolis, C. J., Bartzokas, A., and Katsoulis, B. D. (2002), Spatial and temporal 850 hPa air temperature and sea-surface temperature covariances in the Mediterranean region and their connection to atmospheric circulation. International Journal of Climatology 22: 663–76. Lourens, L. J., Hilgen, F. J., Gudjonsson, L., and Zachariasse, W. J. (1992), Late Pliocene to early Pleistocene astronomically forced sea surface productivity and temperature variations in the Mediterranean. Marine Micropaleontology 19: 49–78. Antonarakou, A., Hilgen, F. J., Van Hoof, A. A. M., Vergnaud-Grazzini, C., and Zachariasse, W. J. (1996), Evaluation of the Plio-Pleistocene astronomical time-scale. Paleoceanography 11: 391–413. Wehausen, R., and Brumsack, H. J. (2001), Geological constraints on tidal dissipation and dynamical ellipticity of the earth over the past three million years. Nature 409: 1029–33. Loye-Pilot, M.-D., Martin, J. M., and Morelli, J. (1989), Atmospheric input of particulate matter and inorganic nitrogen to the northwestern Mediterranean. Water Pollution Research Reports 13: 368–76. McGill, D. A. (1961), A preliminary study of the oxygen and phosphate distribution in the Mediterranean Sea. Deep-Sea Research 8: 259–69. McKenzie, J. A. (1993), Pluvial conditions in the eastern Sahara following the penultimate deglaciation: implications for changes in atmospheric circulation patterns with global warming. Palaeogeography, Palaeoclimatology, Palaeoecology 103: 95–105. Macklin, M. G., Fuller, I. C., Lewin, J., Maas, G. S., Passmore, D. G., Rose, J., Woodward, J. C., Black, S., Hamlin, R. H. B., and Rowan, J. S. (2002), Correlation of fluvial sequences in the Mediterranean basin over the last 200 ka and their relationship to climate change. Quaternary Science Reviews 21: 1633–41. Magri, D. and Parra, I. (2002), Late Quaternary western Mediterranean pollen records and African wind. Earth and Planetary Science Letters 200: 401–8. Maheras, P., Xoplaki, E., Davies, T., Martin-Vide, J., Bariendos, M., and Alcoforado, M. J. (1999), Warm and cold monthly anomalies across the Mediterranean basin and their relation-

ship with circulation; 1860–1990. International Journal of Climatology 19: 1697–1715. Malanotte-Rizzoli, P. and Bergamasco, A. (1989), The circulation of the eastern Mediterranean, Part I. Oceanologica Acta 12: 335–51. (1991), The wind and thermally driven circulation of the eastern Mediterranean Sea. Part II: The baroclinic case. Dynamics Atmososphere Oceans 15: 355–419. and Hecht, A. (1988), Large-scale properties of the eastern Mediterranean: a review. Oceanologica Acta 11: 323–35. Mandel, R. D. and Simmons, A. H. (2001), Prehistoric occupation of Late Quaternary Landscapes near Kharga Oasis, Western Desert of Egypt. Geoarchaeology 16: 95–117. Martinson, D. G., Pisias, N. G., Hayes, J. D., Imbrie, J., Moore, T. C., and Shackleton, N. J. (1987), Age dating and the orbital theory of the ice-ages: development of a high-resolution 0 to 300,000 year chronostratigraphy. Quaternary Research 27: 1–29. Matthews, A., Ayalon, A., and Bar-Matthews, M. (2000), D/H ratios of fluid inclusions of Soreq Cave (Israel) speleothems as a guide to the Eastern Mediterranean Meteoric Line relationships in the last 120 ky. Chemical Geology 166: 183–91. Matthiessen, S. and Haines, K. (2003), A hydraulic box model study of the Mediterranean response to postglacial sea-level rise. Paleoceanography 18: doi 1029/2003PA000880. Mayewski, P. A., Rohling, E. J., Stager, J. C., Karlén, W., Maasch, K., Meeker, L. D., Meyerson, E., Gasse, F., Van Kreveld, S., Holmgren, K., Lee-Thorp, J., Rosqvist, G., Rack, F., Staubwasser, M., Schneider, R. R., Steig, E. (2004), Holocene Climate Variability. Quaternary Research 62: 243–55. MEDATLAS (1997), Mediterranean Hydrographic Atlas (Mast Supporting Initiative; MAS2-CT93- 0074): CD-ROM. MEDOC group (1970), Observation of formation of deep water in the Mediterranean Sea. Nature 277: 1037–40. MEDRIFF consortium (1995), Three brine lakes discovered in the seafloor of the eastern Mediterranean. EOS 76: 313–18. Mercone, D., Thomson, J., Croudace, I. W., Siani, G., Paterne, M., and Troelstra, S. (2000), Duration of S1, the most recent sapropel in the eastern Mediterranean Sea, as indicated by accelerator mass spectrometry radiocarbon and geochemical evidence, Paleoceanography 15: 336–47. Abu-Zied, R. H., Croudace, I. W., and Rohling, E. J. (2001), High-resolution geochemical and micropalaeontological profiling of the most recent eastern Mediterranean sapropel, Marine Geology 177: 25–44. Meyers, P. A. and Negri, A. (2003) (eds.), Paleoclimatic and Paleoceanographic Records in Mediterranean Sapropels and Mesozoic Black Shales. Palaeogeography, Palaeoclimatology, Palaeoecology 190. Middleton, N. J. (1985), Effect of drought on dust production in the Sahel. Nature 409: 431–4. Miller, A. R. (1963), Physical Oceanography of the Mediterranean Sea: A discourse. Rapp. Comm. Int. Mer Médit. 17: 857–71. (1972), Speculations concerning bottom circulation in the Mediterranean Sea, in D. J. Stanley (ed.), The Mediterranean Sea. Dowden, Hutchinson, & Ross, Stroudsburg, Pa. 37–42. (1974), Deep convection in the Aegean Sea. Coll. Int. CNRS 215: 155–63. Tchernia, P., Charnock, H., and McGill, D. A. (1970), Mediterranean Sea Atlas of Temperature, Salinity, Oxygen: Profiles and Data From Cruises of R.V. Atlantis and R.V. Chain, With Distribution of Nutrient Chemical Properties, ed. A. E. Maxwell et al. Alpine, Braintree, Mass.

The Marine Environment: Present and Past Molinaroli, E. (1996), Mineralogical characterisation of Saharan dust with a view to its final destination in Mediterranean sediments, in S. Guerzone and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 153–62. Mommersteeg, H. J. P. M., Loutre, M. F., Young, R., Wijmstra, T. A., and Hooghiemstra, H. (1995), Orbital forced frequencies in the 975,000 year pollen record from Tenaghi Philippon (Greece). Climate Dynamics 11: 4–24. Moreno, A., Cacho, I., Canals, M., Prins, M. A., Sánchez-Goñi, M. F., Grimalt, J. O., and Weltje, G. J. (2002), Saharan dust transport and high-lattitude glacial climatic variability: the Alboran Sea record. Quaternary Research 58: 318–28. Moses, T., Kiladis, G. N., Diaz, H. F., and Barry, R. G. (1987), Characteristics and frequency of reversals in mean sea level pressure in the North Atlantic sector and their relationship to long-term temperature trends. Journal of Climatology 7: 13–30. Moulin, C., Lambert, C. E., Dulac, F., and Dayan, U. (1997), Control of atmospheric export of dust from North Africa by the North Atlantic Oscillation. Nature 387: 691–4. Myers, P. G. (2002), Flux-forced simulations of the paleocirculation of the Mediterranean. Paleoceanography 17: doi 10.1029/2000PA000613. Haines, K., and Rohling, E. J. (1998), Modelling the paleocirculation of the Mediterranean: The last glacial maximum and the Holocene with emphasis on the formation of sapropel S1. Paleoceanography 13: 586–606. Naval Oceanography Command (1987), U.S. Navy Climatic Study of the Mediterranean Sea. Naval Oceanography Command Detachment, Asheville, NC. Nijenhuis, I. A., Schenau, S. J., van der Weijden, C. H., Hilgen, F. J., Lourens, L. J., and Zachariasse, W. J. (1996), On the origin of upper Miocene sapropelites: a case study from the Faneromeni section, Crete (Greece). Paleoceanography 11: 633–45. Nof, D. (1979), On man-induced variations in the circulation of the Mediterranean Sea. Tellus 31: 558–64. Nolet, G. J. and Corliss, B. H. (1990), Benthic foraminiferal evidence for reduced deep water circulation during sapropel formation in the eastern Mediterranean. Marine Geology 94: 109–30. Olausson, E. (1961), Studies of deep-sea cores. Reports of the Swedish Deep Sea Expedition 1947–1948 8: 353–91. O’Neill-Baringer, M. and Price, J. F. (1999), A review of the physical oceanography of the Mediterranean outflow. Marine Geology 155: 63–82. Ozsoy, E. (1981), On the atmospheric factors affecting the Levantine Sea. European Centre for Medium Range Weather Forecasts, Technical Report 25. Pachur, H. J. (2001), Holozäne Klimawechsel in den nördlichen Subtropen. Nova Acta Leopoldina NF88 331: 109–31. and Braun, G. (1980), The paleoclimate of the central Sahara, Libya and the Libyan Desert. Palaeoecology of Africa 12: 351–63. Pälike, H. and Shackleton, N. J. (2000), Constraints on astronomical parameters from the geological record for the last 25 Myr. Earth and Planetary Science Letters 182: 1–14. Parrilla, G., Kinder, T. H., and Preller, R. H. (1986), Deep and intermediate Mediterranean water in the western Alboran Sea. Deep-Sea Research 33: 55–88. Paterne, M., Kallel, N., Labeyrie, L., Vautravers, M., Duplessy, J. C., Rossignol-Strick, M., Cortijo, E., Arnold, M., and Fontugne, M. (1999), Hydrological relationship between the

63

North Atlantic ocean and the Mediterranean during the past 15–75 kyr. Paleoceanography 14: 626–38. Peltier, W. R. (2004), Global glacial isostacy and the surface of the ice-age earth: the ICE-5G (VM2) model and GRACE. Annual Review Earth and Planetary Science 32: 111–49. Perissoratis, C. and Piper, D. J. W. (1992), Age, regional variation and shallowest occurence of S1 sapropel in the northern Aegean Sea. Geomarine Letters 12: 49–53. Perkins, H. and Pistek, P. (1990), Circulation in the Algerian Basin during June 1986. Journal of Geophysical Research 95: 1577–85. Petit-Maire, N. and Guo, Z. (1997), In-phase Holocene climate variations in the present-day desert areas of China and Northern Africa, in J. Meco and N. Petit-Maire, Climates of the Past, Proceedings of the UNESCO-IUGS, June 2–7, 1995, CLIP meeting. Universidad de Las Palmas, Gran Canaria, 137–40. Pinardi, N. and Masetti, E. (2000), Variability of the large scale general circulation of the Mediterranean Sea from observations and modelling: a review. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 153–73. Pisias, N. G. and Moore, T. C. (1981), The evolution of Pleistocene climate: a time series approach. Earth and Planetary Science Letters 52: 450–8. Pistek, P., De Strobel, F., and Montanari, C. (1985), Deep Sea circulation in the Alboran Sea. Journal of Geophysical Research 90: 4969–76. Pollak, M. I. (1951), The sources of deep water of the eastern Mediterranean Sea. Deep-Sea Research 10: 128–52. Poulos, S. E., Drakopoulos, P. G., and Collins M. B. (1997), Seasonal variability in sea surface oceanographic conditions in the Aegean Sea (eastern Mediterranean): an overview. Journal of Marine Systems 13: 225–44. Prell, W. L. (1984), Covariance patterns of foraminiferal ‰18 O: and evaluation of Pliocene ice volume changes near 3.2. million years ago. Science 226: 692–4. and Kutzbach, J. E. (1987), Monsoon variability over the past 150,000 years. Journal of Geophysical Research 92: 8411–25.. Prentice, I. C., Guiot, J., and Harrison, S. P. (1992), Mediterranean vegetation, lake levels and palaeoclimate at the last glacial maximum. Nature 360: 658–60. Price, J. F., O’Neil Baringer, M., Lueck, R. G., Johnson, G. C., Ambar, I., Parilla, G., Cantos, A., Kennelly, M. A., and Sanford, T. B. (1993), Mediterranean outflow mixing and dynamics. Science 259: 1277–82. Reichart, G. J. (1997), Late Quaternary variability of the Arabian Sea monsoon and oxygen minimum zone. Geologica Ultraiectina 154: 174. Reid, J. L. (1979), On the contribution of the Mediterranean Sea outflow to the Norwegian-Greenland Sea, Deep-Sea Research 26A: 1199–223. Reiss, Z., Halicz, E., and Luz, B. (1999). Late-Holocene foraminifera from the SE Levantine Basin, Israel Journal of Earth Science 48: 1–27. Richardson, P. L., McCartney, M. S., and Maillard. C. (1991), A search for meddies in historical data. Dynamics of Atmospheres and Oceans 15: 241–65. Bower, A. S., and Zenk, W. (2000), A census of meddies tracked by floats, Progress in Oceanography 45: 209–50. Richez, C. and Gascard, J. C. (1986), Mediterranean water flows when approaching the Strait of Gibraltar. Deep-Sea Research 135–7.

64

Eelco Rohling et al.

Ritchie, J. C., Eyles, C. H., and Haynes, C. V. (1985), Sediment and pollen evidence for an early to mid-Holocene humid period in the eastern Sahara. Nature 314: 352–5. Robinson, A. R. and Golnaraghi, M. (1994), The physical and dynamical oceanography of the Mediterranean Sea, in P. Malanotte-Rizzoli and A. R. Robinson (eds.), Ocean Processes in Climate Dynamics: Global and Mediterranean Examples. Kluwer, Dordrecht, 255–306. and 35 others (1992), General circulation of the eastern Mediterranean. Earth Science Reviews 32: 285–309. Rodwell, M. J. and Hoskins, B. J. (1996), Monsoons and the dynamics of deserts. Quarterly Journal of the Royal Meteorological Society 122: 1385–404. Roether, W. and Schlitzer, R. (1991), Eastern Mediterranean deep water renewal on the basis of chlorofluoromethane and tritium data. Dynamics Atmosphere and Oceans 15: 333–54. and Well, R. (2001), Oxygen consumption in the Eastern Mediterranean. Deep-Sea Research I 48: 1535–51. Schlosser, P., Kuntz, R., and Weiss, W. (1983), Transient tracer studies of the thermohaline circulation of the Mediterranean. Manuscripts for Mediterranean Circulation, NATO Workshop. Manca, B. B., Klien, B., Bregant, D., Georgopoulos, D., Beitzel, V., Kovacevic, V., and Luchetta, A. (1996), Recent changes in Eastern Mediterranean Deep Waters. Science 271: 333–5. Rognon, P. (1987), Aridification and abrupt climatic events on the Saharan northern and southern margins, 20,000 years BP to present, in W. H. Berger and L. D. Labeyrie (eds.), Abrupt Climatic Change: Evidence and Implications. D. Reidel, Dordrecht, 209–20. Rohling, E. J. (1991a ), A simple two-layered model for shoaling of the eastern Mediterranean pycnocline due to glacio-eustatic sea level lowering. Paleoceanography 6: 537–41. (1991b), Shoaling of the eastern Mediterranean pycnocline due to reduction of excess evaporation: implications for sapropel formation. Paleoceanography 6: 747–53. (1994), Review and new aspects concerning the formation of Mediterranean sapropels. Marine Geology 122: 1–28. (1997), Mutual influencing between the Atlantic Ocean and the Mediterranean Sea during the Quaternary, in J. Meco and N. Petit-Maire, Proceedings of the IUGS-UNESCO CLIP IV Meeting, Canary Islands. Universidad de Las Palmas, Gran Canaria, 141–9. (1999a ), Environmental controls on salinity and ‰18 O in the Mediterranean. Paleoceanography 14: 706–15. (1999b) (ed.), Mediterranean paleoclimate and sapropel studies, Marine Geology 153. and Bryden, H. L. (1992), Man-induced salinity and temperature increases in western Mediterranean Deep Water. Journal of Geophysical Research 97: 11191–8. (1994), Estimating past changes in the eastern Mediterranean freshwater budget, using reconstructions of sea level and hydrography. Proceedings Koninklijke Nederlandse Akademie, Ser.B. 97: 201–17. and De Rijk. S. (1999a ), The Holocene Climate Optimum and Last Glacial Maximum in the Mediterranean: the marine oxygen isotope record. Marine Geology 153: 57–75. (1999b), Erratum to ‘The Holocene Climate Optimum and Last Glacial Maximum in the Mediterranean: the marine oxygen isotope record’. Marine Geology 161: 385–7.

and Gieskes, W. W. C. (1989), Late Quaternary changes in Mediterranean Intermediate Water density and formation rate. Paleoceanography 4: 531–45. and Hilgen, F. J. (1991), The eastern Mediterranean climate at times of sapropel formation: a review. Geologie en Mijnbouw 70: 253–64. and Pälike, H. (2005), Centennial-scale climate deterioration with a superimposed cold event around 8,200 years ago. Nature 434: 975–9. and Zachariasse, W. J. (1996), Red Sea outflow during the Last Glacial Maximum. Quaternary International 31: 77–83. Jorissen, F. J., Vergnaud-Grazzini, C., and Zachariasse, W. J. (1993a ), Northern Levantine and Adriatic planktonic foraminifera: reconstruction of paleoenvironmental gradients. Marine Micropaleontology 21: 191–218. Vergnaud-Grazzini, C., and Zaalberg, R. (1993b), Temporary repopulation by low-oxygen tolerant benthic foraminifera within an Upper Pliocene sapropel: Evidence for the role of oxygen depletion in the formation of sapropels. Marine Micropaleontology 22: 207–19. Den Dulk, M., Pujol, C., and Vergnaud-Grazzini, C. (1995), Abrupt hydrographic change in the Alboran Sea (western Mediterranean) around 8000 yrs BP. Deep-Sea Research 42: 1609–19. Jorissen, F. J., and De Stigter, H. C. (1997), 200 Year interruption of Holocene sapropel formation in the Adriatic Sea. Journal of Micropalaeontology 16: 97–108. Fenton, M., Jorissen, F. J., Bertrand, P., Ganssen, G., and Caulet, J. P. (1998a ), Magnitudes of sea-level lowstands of the past 500,000 years. Nature 394: 162–5. Hayes, A., De Rijk, S., Kroon, D., Zachariasse, W. J., and Eisma, D. (1998b), Abrupt cold spells in the northwest Mediterranean. Paleoceanography 13: 316–22. Mayewski, P. A., Hayes, A., Abu-Zied, R. H., and Casford, J. S. L. (2002a ), Holocene atmosphere-ocean interactions: records from Greenland and the Aegean Sea. Climate Dynamics 18: 587–93. Cane, T. R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K. C., Schiebel, R., Kroon, D., Jorissen, F. J., Lorre, A., and Kemp, A. E. S. (2002b), African monsoon variability during the previous interglacial maximum. Earth and Planetary Science Letters 202: 61–75. Mayewski, P. A., and Challenor, P. (2003), On the timing and mechanism of millennial-scale climate variability during the last glacial cycle. Climate Dynamics 20: 257–67. Sprovieri, M., Cane, T. R., Casford, J. S. L., Cooke, S., Bouloubassi, I., Emeis, K. C., Schiebel, R., Rogerson, M., Hayes, A., Jorissen, F. J., and Kroon, D. (2004), Reconstructing past planktic foraminiferal habitats using stable isotope data: a case history for Mediterranean sapropel S5. Marine Micropaleontology 50: 89–123. Rose, J., Meng, X., and Watson, C. (1999), Palaeoclimate and palaeoenvironment responses in the western Mediterranean over the last 140,000 ka: evidence from Mallorca, Spain. Journal of the Geological Society 156: 435–48. Rossignol-Strick, M. (1983), African monsoons, an immediate climate response to orbital insolation. Nature 30: 446–9. (1985), Mediterranean Quaternary sapropels, an immediate response of the African monsoon to variations of insolation. Palaeogeography, Palaeoclimatology, Palaeoecology 49: 237–63.

The Marine Environment: Present and Past (1987), Rainy periods and bottom water stagnation initiating brine accumulation and metal concentrations: 1. the Late Quaternary. Paleoceanography 2: 333–60. (1995), Sea-land correlation of pollen records in the eastern Mediterranean for the glacial-interglacial transition: biostratigraphy versus radiometric time-scale. Quaternary Science Reviews 14: 893–915. Nesteroff, V., Olive, P., and Vergnaud-Grazzini, C. (1982), After the deluge; Mediterranean stagnation and sapropel formation. Nature 295: 105–10. Roussenov, V., Stanev, E., Artale, V., and Pinardi, N. (1995), A seasonal model of the Mediterranean Sea general circulation. Journal of Geophysical Research 100: 13515–38. Ruddiman, W. F. and Kutzbach, J. E. (1989), Forcing of Late Cenozoic northern hemisphere climate by plateau uplift in southern Asia and the American West. Journal of Geophysical Research 94: 18409–27. Raymo, M., and McIntyre, A. (1986), Matuyama 41,000year cycles: North Atlantic Ocean and northern hemisphere ice sheets. Earth and Planetary Science Letters 80: 117–29. McIntyre, A., and Raymo, M. (1987), Paleoenvironmental results from North Atlantic sites 607 and 609, in W. F. Ruddiman et al., Initial Reports Deep Sea Drilling Project 94: 855–78. Martinson, D., Clement, B., and Backman, J. (1989), Pleistocene evolution: northern hemisphere ice sheets and North Atlantic Ocean. Paleoceanography 4: 353–412. Rumney, G. R. (1968), Climatology and the World’s Climates. Macmillan, New York. Saaroni, H., Bitan, A., Alpert, P., and Ziv, B. (1996), Continental Polar outbreaks into the Levant and Eastern Mediterranean. International Journal of Climatology 16: 1175–91. Said, R. (1981), The Geological Evolution of the River Nile. Springer, New York. Samuel, S., Haines, K., Josey, S., and Myers, P. G. (1999), Response of the Mediterranean Sea thermohaline circulation to observed changes in the winter wind stress field in the period 1980–1993. Journal of Geophysical Research 104: 7771–84. Sánchez-Goñi, M. F., Sierro, F. J., Peypouquet, J. P., Grimalt, J. O., Shackleton, N. J., Cacho, I., Turon, J. L., and Guiot, J. (2002), Synchroneity between marine and terrestrial responses to millennial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19: 95–105. Schlitzer, R., Roether, W., Oster, H., Junghans, H. G., Hausmann, M., Johannsen, H., and Michelato, A. (1991), Chlorofluoromethane and oxygen in the eastern Mediterranean. Deep-Sea Research 38: 1531–51. Schmiedl, G., Mitschele, A., Beck, S., Emeis, K. C., Hemleben, Ch., Schultz, H., Sperling, M., and Weldeab, S. (2003), Benthic foraminiferal record of ecosystem variability in the eastern Mediterranean Sea during times of sapropel S5 and S6 deposition. Palaeogeography, Palaeoclimatology, Palaeoecology 190: 139–64. Scientific Staff Cruise BAN84 (1985), Gypsum precipitation from cold brines in an anoxic basin in the eastern Mediterranean. Nature 314: 152–4. Scrivner, A., Vance, D., and Rohling, E. J. (2004), New neodymium isotope data quantify Nile involvement in Mediterranean anoxic episodes. Geology 32: 565–8. Shackleton, N. J. and Opdyke, N. D. (1973), Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28238: Oxygen isotope temperatures and ice volume on a 105 and 106 year time scale. Quaternary Research 3: 39–55.

65

(1976), Oxygen isotope and paleomagnetic stratigraphy of equatorial Pacific core V28-238, late Pliocene to latest Pleistocene, in: R. M. Cline and J. D. Hays (eds.) Investigations of Late Quaternary Palaeoceanography and Palaeoclimatology. Memoirs of the Geological Society of America 145: 449–64. (1977), Oxygen isotope and paleomagnetic evidence for early northern hemisphere glaciation. Nature 270: 216–19. Sharaf el Din, S. H. (1977), Effect of Aswan High Dam on the Nile flood and on the estuarine and coastal circulation pattern along the Mediterranean Egyptian coast. Limnology and Oceanography 22: 194–207. Shaw, H. F. and Evans, G. (1984), The nature, distribution and origin of a sapropelic layer in sediments of the Cilicia Basin, northeastern Mediterranean. Marine Geology 61: 1–12. Shimmield, G. B., Mowbray, S. R., and Weedon, G. P. (1990), A 350-ka history of the Indian Southwest monsoon—evidence from deep-sea cores, northwest Arabian Sea. Transactions of the Royal Society, Edinburgh, Earth Science 81: 289–99. Siddall, M., Rohling, E. J., Almogi-Labin, A., Hemleben, Ch., Meischner, D., Schmelzer, I., and Smeed, D. A. (2003), Sea-level fluctuations during the last glacial cycle. Nature 423: 853–8. Stanley, D. J. (1978), Ionian Sea sapropel distribution and late Quaternary palaeoceanography in the eastern Mediterranean. Nature 274: 149–52. (1996), Nile Delta: extreme case of sediment entrapment on a delta plain and consequent coastal land loss. Marine Geology 129: 189–95. and Wingerath, J. (1996), Clay mineral distributions to interpret Nile cell provenance and dispersal: I. Lower River Nile to delta. Journal of Coastal Research 12: 911–29. Nir, Y., and Galili, E. (1998), Clay mineral distributions to interpret Nile cell provenance and dispersal: II. Offshore margin between Nile to delta and northern Israel. Journal of Coastal Research 14: 196–217. Stommel, H., Bryden, H., and Mangelsdorf, P. (1973), Does some of the Mediterranean outflow come from great depth? Pure and Applied Geophysics 105: 879–89. Stratford, K., Williams, R. G., and Myers, P. G. (2000), Impact of the circulation on sapropel formation in the eastern Mediterranean. Global Biogeochemical Cycles 14: 683–95. Suc, J. P. (1984), Origin and evolution of the Mediterranean vegetation and climate in Europe. Nature 307: 429–32. Szabo, B. J., Haynes Jr., C. V., and Maxwell, T. A. (1995), Ages of Quaternary pluvial episodes determined by uranium-series and radiocarbon dating of lacustrine deposits of Eastern Sahara. Palaeogeography, Palaeoclimatology, Palaeoecology 113: 227– 42. Tang, C. M. and Stott, L. D. (1993), Seasonal salinity changes during Mediterranean sapropel deposition 9000 years B.P.: evidence from isotopic analyses of individual planktonic foraminifera. Paleoceanography 8: 473–93. Theocharis, A., (1989), Deep water formation and circulation in the Aegean Sea. Reports in Meteorology and Oceanography 40/1: 335–59. and Georgopoulos, D. (1993), Dense water formation over the Samothraki and Limnos Plateaux in the north Aegean Sea (eastern Mediterranean Sea). Continental Shelf Research 13: 919–39. Thomson, J., Higgs, N. C., Wilson, T. R. S., Croudace, I. W., De Lange, G. J., and Van Santvoort, P. J. M. (1995), Redistribution and geochemical behaviour of redox-sensitive elements around

66

Eelco Rohling et al.

S1, the most recent eastern Mediterranean sapropel. Geochimica et Cosmochimica Acta 59: 3487–501. Mercone, D., de Lange, G. J., and van Santvoort, P. J. M. (1999), Review of recent advances in the interpretation of eastern Mediterranean sapropel S1 from geochemical evidence. Marine Geology 153: 77–89. Thunell, R. C. (1986), Pliocene-Pleistocene climatic changes: evidence from land-based and deep-sea marine records. Mem. Societa Geologica Italiana 31: 135–43. and Williams, D. F. (1983), Paleotemperature and paleosalinity history of the eastern Mediterranean during the late Quaternary. Palaeogeography, Palaeoclimatology, Palaeoecology 44: 23–39. (1989), Glacial-Holocene salinity changes in the Mediterranean Sea: hydrographic and depositional effects. Nature 338: 493–6. and Kennett, J. P. (1977), Late Quaternary paleoclimatology, stratigraphy and sapropel history in eastern Mediterranean deep-sea sediments. Marine Micropaleontology 2: 371–88. and Cita, M. B. (1983), Glacial anoxia in the eastern Mediterranean. Journal of Foraminiferal Research 13: 283–90. and Howell, M. (1987), Atlantic-Mediterranean water exchange during the late Neogene. Paleoceanography 2: 661–78. Rio, D., Sprovieri, R., and Vergnaud-Grazzini, C. (1991), An overview of the post-Messinian paleoenvironmental history of the western Mediterranean. Paleoceanography 6: 143–64. Tintoré, J. D., La Violette, P. E., Blade, I., and Cruzado, A. (1988), A study of an intense density front in the eastern Alboran Sea: The Almeria-Oran Front. Journal of Physical Oceanography 18: 1384–97. Trewartha, G. T. (1966), The Earth’s Problem Climates. Methuen, London. Trigo, I. F., Davies, T. D., and Bigg, G. R. (1999), Objective climatology in the Mediterranean region. Journal of Climate 12: 1685–96. – (2000), Decline in Mediterranean rainfall caused by weakening of Mediterranean cyclones. Geophysical Research Letters 27: 2913–16. Troelstra, S. R. (1987), Late Quaternary sedimentation in Tyro and Kretheus Basins, southeast of Crete. Marine Geology 75: 77–91. Tzedakis, P. C. (1993), Long-term tree populations in northwest Greece through multiple climatic cycles. Nature 364: 437–40. (1999), The last climatic cycle at Kopais, central Greece. Journal of the Geological Society, London 156: 425–34. Frogley, M. R., Lawson, I. T., Preece, R. C., Cacho, I., and de Abreu, L. (2004), Ecological thresholds and patterns of millennial-scale climate variability: the response of vegetation in Greece during the last glacial period. Geology 32: 109–12. UNDP/UNESCO (1978), Coastal Protection Studies. Project Findings and Recommendations. Paris, UNDP/EGY/73/063. Van Os, B. J. H., Lourens, L. J., Hilgen, F. J., De Lange, G. J., and Beaufort, L. (1994), The formation of Pliocene sapropels and carbonate cycles in the Mediterranean: diagenesis, dilution, and productivity. Paleoceanography 9: 601–17. van Straaten, L. M. J. U. (1966), Micro-malacological investigation of cores from the southeastern Adriatic Sea. Proceedings, Koninklijke Nederlandse Akademie Sciences, Serie B 69: 429–45.

(1970), Holocene and late-Pleistocene sedimentation in the Adriatic Sea. Geologische Rundschau 60: 106–31. (1972), Holocene stages of oxygen depletion in deep waters of the Adriatic Sea, in D. J. Stanley (ed.), The Mediterranean Sea. Dowden, Hutchingson, & Ross, Stroudsburg, Pa., 631–43. (1985), Molluscs and sedimentation in the Adriatic Sea during late-Pleistocene and Holocene times. Giornale Geologia, Serie 3a 47: 181–202. Vergnaud-Grazzini, C. (1985), Mediterranean late Cenozoic stable isotope record: stratigraphic and paleoclimatic implications, in D. J. Stanley and F. C. Wezel (eds.), Geological Evolution of the Mediterranean Basin. Springer, New York, 413–51. Ryan, W. B. F., and Cita, M. B. (1977), Stable isotope fractionation, climatic change and episodic stagnation in the eastern Mediterranean during the late Quaternary. Marine Micropaleontology 2: 353–70. Borsetti, A. M., Cati, F., Colantoni, P., d’Onofrio, S., — Saliège, J. F., Sartori, R., and Tampieri, R. (1988), Palaeoceanographic record of the last deglaciation in the Strait of Sicily. Marine Micropaleontology 13: 1–21. Capotondi, L., and Lourens, L. J. (1993), A refined Pliocene to early Pleistocene chronostratigraphic frame at ODP Hole 653A (West Mediterranean), Marine Geology 117: 329–49. Verhallen, P. J. J. M. (1991), Late Pliocene to Early Pleistocene Mediterranean mud-dwelling foraminifera; influence of a changing environment on community structure and evolution. Utrecht Micropaleontological Bulletin 40. Voelker, A. H., et al. (2002), Global distribution of centennialscale records for marine isotope stage (MIS) 3: a database. Quaternary Science Reviews 21: 1185–212. Wahby, S. D. and Bishara, N. F. (1981), The effect of the river Nile on Mediterranean water, before and after the construction of the High Dam at Aswan, in J. M. Martin, J. D. Burton, and D. Eisma (eds.), River Inputs to Ocean Systems. UNEP-UNESCO, Switzerland, 311–18. Wallmann, K., Suess, E., Westbrook, G. H., Winkler, G., Cita, M. B., and the MEDRIFF consortium (1997), Salty brines on the Mediterranean sea floor. Nature 387: 31–2. Wehausen, R. and Brumsack H. J. (2000), Chemical cycles in Pliocene sapropel-bearing and sapropel-barren eastern Mediterranean sediments. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 325–52. Weldeab, S., Emeis, K. C., Hemleben, C., and Siebel, W. (2002), Provenance of lithogenic surface sediments and pathways of riverine suspended matter in the Eastern Mediterranean Sea: evidence from 143 Nd/144 Nd and 87 Sr/86 Sr ratios. Chemical Geology 186: 139–49. Wijmstra, T. A., Young, R., and Witte, H. J. L. (1990), An evaluation of the climatic conditions during the Late Quaternary in northern Greece by means of multivariate analysis of palynological data and comparison with recent phytosociological and climatic data. Geologie en Mijnbouw 69: 243–51. Williams, D. F., Thunell, R. C., and Kennett, J. P. (1978), Periodic fresh-water flooding and stagnation of the eastern Mediterranean during the late Quaternary. Science 201: 252–4. Williams, M. A. J., Adamson, D. A., Cock, B., and McEvedy, R. (2000), Late Quaternary environments in the White Nile region, Sudan. Global and Planetary Change 26: 305–16. Wüst, G. (1960), Die tiefenzirkulation des Mittelländischen Meeres in den Kernschichten des Zwischen-und des Tiefenwassers. Deutsche Hydrographische Zeitschrift 13: 105–31.

The Marine Environment: Present and Past Wüst, G. (1961), On the vertical circulation of the Mediterranean Sea, Journal of Geophysical Research 66: 3261–71. Yüce, H. (1995), Northern Aegean water masses. Estuarine, Coastal and Shelf Science 41: 325–43. Zachariasse, W. J. and Spaak, P. (1983), Middle Miocene to Pliocene paleoenvironmental reconstruction of the Mediterranean and adjacent Atlantic Ocean: planktonic foraminiferal record of southern Italy, in J. E. Meulenkamp (ed.) Reconstruction of marine paleoenvironments. Utrecht Micropaleontological Bulletin 30: 91–110.

67

Zijderveld, J. D. A., Langereis, C. G., Hilgen, F. J., and Verhallen, P. J. J. M. (1989), Early Late Pliocene biochronology and surface water temperature variations in the Mediterranean. Marine Micropaleontology 14: 339–55. Gudjonsson, L., Hilgen, F. J., Langereis, C. G., Lourens, L. J., Verhallen, P. J. J. M., and Zijderveld, J. D. A. (1990), Late Gauss to Early Matuyama invasions of Neogloboquadrina atlantica in the Mediterranean and associated record of climatic change. Paleoceanography 5: 239–52.

This chapter should be cited as follows Rohling, E. J., Abu-Zied, R. H., Casford, J. S. L., Hayes, A., and Hoogakker, B. A. A. (2009), The marine environment: present and past, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 33–67.

This page intentionally left blank

3

The Climate System Andrew Harding, Jean Palutikof, and Tom Holt

Introduction The Mediterranean region has a highly distinctive climate due to its position between 30 and 45◦ N to the west of the Euro-Asian landmass. With respect to the global atmospheric system, it lies between subtropical high pressure systems to the south, and westerly wind belts to the north. In winter, as these systems move equatorward, the Mediterranean basin lies under the influence of, and is exposed to, the westerly wind belt, and the weather is wet and mild. In the summer, as shown in Figure 3.1, the Mediterranean lies under subtropical high pressure systems, and conditions are hot and dry, with an absolute drought that may persist for more than two or three months in drier regions. Climates such as this are relatively rare, and the Mediterranean shares its winter wet/summer dry conditions with locations as distant as central Chile, the southern tip of Cape Province in South Africa, southwest Australia in the Southern Hemisphere, and central California in the Northern Hemisphere. All have in common their mid-latitude position, between subtropical high pressure systems and westerly wind belts. They all lie on the westerly side of continents so that, in winter, when the westerly wind belts dominate over their locations, they are exposed to rain-bearing winds. In the Köppen classification (Köppen 1936), these climates are known as Mediterranean (Type Cs, which is subdivided in turn into maritime Csb and continental Csa). The influence of the Mediterranean Sea means that the Mediterranean-type climate of the region extends much further into the continental landmass than elsewhere, and is not restricted to a narrow ocean-facing strip. Nevertheless, within the Mediterranean region

climate is modified by position and topographic influences can be important. The proximity of the western Mediterranean to the Atlantic Ocean gives its climate a maritime flavour, with higher rainfall and milder temperatures throughout the year. The eastern Mediterranean lies closer to the truly continental influences of central Europe and Asia. Its climate is drier, and temperatures are hotter in summer and colder in winter than in the west. Annual rainfall is typically around 750 mm in Rome, but only around 400 mm in Athens (Figures 3.2 and 3.3a). The southern shore of the Mediterranean Sea is drier and hotter than the northern shore. In the extreme south-east, which lies permanently under the subtropical high pressure belts, the climate becomes arid and hot, and no longer falls into the Köppen Mediterranean class. Here the Sahara desert meets the Mediterranean coast and mean annual rainfall in Alexandria is just 178 mm, with five absolutely dry months each year from May to September (Figures 3.2 and 3.3b). Winter and summer mean temperature and mean monthly precipitation are shown for five locations (including Lisbon on the Atlantic seaboard) in the region in Figure 3.2. Annual rainfall across the Mediterranean region is shown in Figure 3.3a with a general decrease in values moving to the south and east although this is modified by topography. In the high mountains of the Mediterranean the climate can be very humid indeed with annual precipitation exceeding 3,000 mm (Chapter 12). Altitude also plays an important role and climate can change dramatically moving inland from the coast. Snow and freezing conditions are rare in low-lying areas, but in mountainous regions such as the Apennines of Italy and Pindus of Greece winters can be severe,

70

Andrew Harding, Jean Palutikof, and Tom Holt

January

July

H 1020 H

40°N

40°N

102

0

L

H

20 10

1000

L

MF

H

L

H 20°N

1010

20°N L

ITF

1015 L

1010

ITF 20°W L 0°

H

0° 20°W

1010

20°E

H

40°E

40°E 0°

Monthly precipitation (mm)

20°E

50–400 >400

Fig. 3.1. The location of the Mediterranean region in relation to the large-scale atmospheric circulation (modified from Barry and Chorley 1992). See Figure 2.4.

and thick snow is common, with glaciers present in both the Pyrenees and Apennines (Chapter 12). Indeed, there are places in the Mediterranean where tourists can spend the morning skiing and the afternoon on the beach, including ski resorts in the Troodos Mountains of Cyprus, and Faraya-Mzaar in Lebanon. Under these conditions, the critical climatic factor for the region is rainfall (Figure 3.3). Although people may feel that winters are too cold, and that in summer the heat becomes excessive, in fact temperatures are rarely a limiting factor to activities. But rainfall, or rather lack of it, may be. This becomes a particular problem on the many small inhabited islands of the Mediterranean, many of which are magnets for tourists in the summer, when there is no rain. Conflicts arise between requirements for domestic supply, irrigation, and tourism (Chapter 21). The water supply comes mainly from groundwater, and salinization and mineralization of the supply is common (Burak et al. 2004; Zalidis et al. 2002). Malta has a population of around 394,000, which triples each summer with the influx of tourists, who provide the main economic activity of the island. Average rainfall is 578 mm per annum. The country has no permanent rivers or lakes, and the water supply has traditionally been from the groundwater, which has become inadequate to support tourism. Malta has implemented a programme to desalinate sea water, and up to 70 per cent of Malta’s water now comes from desalination plants (Bremere et al. 2001).

Characteristics of the Present-day Climate In the western Mediterranean, the weather is principally dictated by proximity to the Atlantic Ocean and temperatures are less extreme than in the eastern Mediterranean. The highest daytime temperature recorded in Almeria, on the south-eastern coast of Spain, is 38◦ C, compared with 43◦ C in Athens. The lowest night time temperature recorded in Almeria is 0◦ C, compared with −4◦ C in Athens. Throughout the Mediterranean, winters are wet and summers are dry. Typically, between 70 and 80 per cent of rainfall is received between October and March, and around 40 per cent between December and February. Conditions become more extreme in the south-eastern Mediterranean and, in Alexandria, 98 per cent of rainfall is received between October and March, and 70 per cent between December and March (Figure 3.3b). Rainfall in the Mediterranean occurs in association with depressions, or centres of low pressure. As in middle and high latitudes, the development and steering of individual depressions is associated with the general thermal pattern and winds in the upper troposphere. The majority of depressions in the Mediterranean have their origins as lee depressions forming, for example, in the lee of the Alps and Atlas mountains (HMSO 1962). However, other mechanisms operate. In the Iberian Peninsula, it has been estimated that around half of rain-producing depressions are of Atlantic origin. These

The Climate System

71

Fig. 3.2. Seasonal temperature and rainfall variations at selected sites around the Mediterranean.

Atlantic depressions weaken as they track across the peninsula and seldom bring rain to the Mediterranean basin. However, they may trigger the formation of new cyclones within the basin itself (Trigo et al. 1999, 2002). When waves of cold air, which accompany Atlantic depressions, encounter warm moist air over the

Mediterranean, the accompanying vertical instability, which may be enhanced in the presence of mountains, leads to the development of vigorous depressions, high rainfall, and high wind speeds (HMSO 1962). Thermal lows can also be important, for example, over the Iberian Peninsula (Chapter 18).

72

Andrew Harding, Jean Palutikof, and Tom Holt

(a)

1500 1000 750 500

Annual rainfall (mm/year)

250 0

0

500 km

0

500 km

(b)

7 5 3

Number of dry months

1 0 (in the north)

Fig. 3.3. (a) Mean annual rainfall and (b) length of the dry season across the Mediterranean basin (modified from Grenon and Batisse 1989).

The Climate System

The Influence of the Global Circulation Jet-streams in the Upper Atmosphere Jet streams are narrow, fast-flowing streams of air at about 11 km above the earth’s surface. Two westerly flowing jet streams can affect the Mediterranean region: the polar-front, most commonly found between latitudes 30 and 70◦ N, and the subtropical jet, between latitudes 20 and 50◦ N. Both are westerly flowing but follow a meandering course. Because they mark the boundary between cooler and warmer air masses, and because they can steer underlying surface depressions, their position at any point in time is important in determining the character of the surface weather. The polar-front jet stream generally lies north of the Mediterranean, but in winter it may be present and is then associated with outbreaks of cold air from the north (HMSO 1962). Abnormal summer heatwaves have been associated with a subtropical jet to the north of its normal position (Balafoutis and Makrogiannis 2001; Baldi et al. 2006).

Large-scale Pressure Patterns The Mediterranean is influenced by planetary-scale pressure patters that shift seasonally, moving equatorward in winter and poleward in summer as shown in Figure 3.1. This brings the Mediterranean basin under the influence of the subtropical high pressure belts in summer and under the influence of the westerly wind belts in winter. These large-scale synoptic controls are also discussed in relation to the nature of the marine environment in Chapter 2. This pattern dictates the large-scale average climate conditions which affect the Mediterranean. Superimposed on these planetary-scale average patterns, the behaviour of pressure centres over the Atlantic Ocean, and the large landmasses of Europe, Asia, and Africa influence the inter-annual and intra-seasonal variability of the regional climate. The behaviour of these pressure centres and their relationship with neighbouring centres (which may involve a see-sawing or oscillation of high pressure in one location and low pressure in another, known as a teleconnection pattern), is captured through the calculation of pressure indices. Some of the principal pressure indices and oscillations affecting the Mediterranean are described below.

73

The North Atlantic Oscillation (NAO) has been found to influence rainfall amounts, especially in winter, in the north-western Mediterranean (Hurrell 1995) and, more remotely, in the eastern Mediterranean (Eshel and Farrell 2000; Eshel et al. 2000). The NAO is a measure of the track of storms and depressions across the North Atlantic Ocean and into Europe. A high NAO winter is one when the storm track is strong, with a north-eastward orientation taking depressions into north-west Europe. A low NAO winter has a weaker storm track with an east–west orientation taking depressions into Mediterranean Europe. The north-western Mediterranean will tend to have wetter winters when the NAO is weak and drier winters when the NAO is strong (Hurrell 1995). The eastern Mediterranean lies under continental influences, accounting for conditions more extreme than experienced in the west. The Indian Monsoon also plays a role in ensuring that the eastern Mediterranean is exceptionally dry in summer. The heating and associated uplift of air over India during the monsoon is linked to subsiding air and hence dry conditions in the eastern Mediterranean (Raicich et al. 2003). In the summer of 2002, Indian Monsoon rainfall was much below normal, and there was exceptionally heavy rain in many parts of southern and central Europe. It has been suggested that El Niño and La Niña events may affect Mediterranean climates. The Walker circulation is an east–west atmospheric circulation pattern of rising air above Indonesia and the western Pacific and sinking air above the eastern Pacific. Under normal conditions, Indonesia is wet and the eastern Pacific is dry. During El Niño events there is a weakening of the Walker circulation. During La Niña events the Walker circulation is especially strong. The effects of an El Niño period are directly linked to areas within the Pacific basin, causing droughts over Indonesia and intense rainfall events, flooding, and landslides in Peru. More importantly, its influence on seasonal weather conditions has been detected as far away as India, Africa, Antarctica, and North America (Bromwich et al. 2000). In the eastern Mediterranean there is some evidence that El Niño events are positively correlated with winter rainfall (Kadioglu et al. 1999; Price et al. 1998; Rodo et al. 1997). Other relationships between Mediterranean weather conditions and El Niño/La Niña events have been suggested, but these are generally weak or impermanent. Kutiel et al. (2002) found a teleconnection between pressure centres over the North Sea and the Caspian Sea. The North Sea Caspian Pattern Index (NCPI) is

74

Andrew Harding, Jean Palutikof, and Tom Holt

calculated from the North Sea–Caspian Sea pressure difference. Between October and April, when the NCPI is positive, temperatures are high over the Balkans, western Turkey, and the Middle East, with below average precipitation. A negative index during an extended winter period is associated with low temperatures and above-average precipitation in this region. Kutiel and Benaroch (2002) found that this index was a better predictor of the climate over the eastern Mediterranean region than the NAO.

The Mediterranean Oscillation So far we have considered the influence of pressure centres outside the Mediterranean basin. However, the elongated west–east shape of the Mediterranean Sea, and its size in relation to the global circulation, have led a number of authors to suggest that pressure differences between the western and eastern ends of the basin may influence climate variability at inter-annual scales. This teleconnection has been called the Mediterranean Oscillation and the Mediterranean Pressure Index (MPI). Here we call it the Mediterranean Oscillation (MO). It has variously been calculated from the pressure difference between Algiers and Cairo (e.g. Corte-Real et al. 1998), between Gibraltar and Israel (Lod Airport) (e.g. Palutikof 2003), and between Mersa Matruh (Egypt) and Marseilles (France) (e.g. Raicich et al. 2003). A number of different influences have been attributed to the MO/MPI. Maheras and Kutiel (1999) found that favourable circulation for high temperatures in the western basin (southerly flow) is associated with unfavourable circulation in the eastern basin (northerly flow) and vice versa. Piervitali et al. (1999) found high negative correlations between eastern Iberian and Italian rainfall and the MO/MPI, exceeding the relationship with the NAO throughout the year. Palutikof (2003) found significant negative relationships between the MO/MPI and rainfall over the western and central Mediterranean in winter and these were higher than the relationship between the NAO and rainfall in these locations. Strongly significant negative relationships between the MO/MPI and autumn rainfall were also found in the western basin. However, no relationship was found between the MO/MPI and rainfall in the eastern Mediterranean. Raicich et al. (2003) found a significant correlation between their MPI and the north–south wind component in the eastern Mediterranean (the Etesian wind regime), and hence negative correlations with Sahelian rainfall and the Indian monsoon.

Areas of Cyclogenesis In some preferred locations, interactions between the large-scale circulation and the complex geography create conditions ideal for the formation of cyclones. These locations, and the characteristics of the cyclones they produce, are described below and summarized in Table 3.1. The Mediterranean Sea is one of the most active regions of cyclogenesis in the world. Although systems are generally weak and shallow, occasionally strong, fast-moving, and active low pressure systems form, bringing very disturbed weather and severe conditions to the region. Strong cyclones may be associated with rainfall of up to 200 mm in 24 hours (Mahovic et al. 2005), and up to 800 mm in 24 hours has been recorded (Peñarrocha et al. 2002). Deep Mediterranean cyclones are commonly associated with high wind speeds, as shown in Chapter 18.

Cyclogenesis in the Northern Mediterranean Cyclones form in three principal areas of the northern Mediterranean: the Gulf of Genoa, the Aegean Sea, and the Black Sea (Figure 3.4). The cyclones produced are generally sub-synoptic in scale. They are triggered by the passage of remnant North Atlantic synoptic systems and their interactions with local topography. Cyclones may therefore form consecutively at the three centres as a single North Atlantic system passes each in turn (Trigo et al. 2002). Of these three centres, the best known is the Gulf of Genoa. Genoa cyclones, although most frequent in winter, may form throughout the year. They generally remain stationary to the south of the Alps. If they do move, they will follow one of two tracks. First, if there are anticyclonic conditions over the Balkans, Turkey, and the Black Sea, cyclones will move out of the Gulf of Genoa in a south-easterly direction towards the Ionian Sea. This provides ideal conditions for a Bora wind to develop as discussed below. Second, cyclones may move north-easterly across the Alps, and may bring extensive rainfall and catastrophic flooding to Austria, Germany, the Czech Republic, and Poland, as in the summer of 2002. Such a path may be associated with the development of Sirocco winds if the circulation of the low extends southwards into North Africa allowing air from the desert to move north. This second path is more commonly followed if cyclogenesis is

The Climate System

75

TABLE 3.1. Centres of cyclogenesis in the Mediterranean Region The Adriatic and Ligurian Seas, including Gulf of Genoa Iberian Peninsula North Africa: Sahara Aegean Sea

Eastern Black Sea Cyprus

Middle East

Associated track

Associated mechanism

Seasonality

• SE-ward direction (Italy, Albania, Greece) • NE from the Adriatic into the Balkans Quasi-stationary

Lee-effect cyclogenesis and conditional instability, upper level vorticity

Declines in intensity towards summer

Creates intense rain across Buzzi and Tibaldi a large sector of the 1978 western basin. Associated with the Mistral

Thermally induced

Peaks June–August

Usually dry but may be thunderstorms Source of important spring rainfall, transported dust, and the Scirocco Increased storminess

• NE into the Med. Lee-effect cyclogenesis due • E along the African to Atlantic flow and coast towards Greece thermal instability NE towards the Black Conditional instability, Sea regenerated Genoan cyclones, lee-effect cyclogenesis NE into Europe Conditional instability, Aegean cyclogenesis E into the Middle East Lee-effect reintensification of western depressions

E and NE into Asia

Asian Monsoon mechanisms

Peaks May–June

Impacts

Declines in intensity towards summer strongest in January Declines in intensity Contributes to annual towards summer rainfall peak Increases in intensity Important source of rain towards summer (but also storms) for Cyprus, southern Turkey, and the Middle East Increases in intensity Dry and settled weather towards summer

References

Alonso et al. 1994 Egger et al. 1995

Flocas and Karacostas 1996 Radinovic 1987 Barry and Chorley 1992; Lagouvardos et al. 1996 Barry and Chorley 1992

Source: Harding (2006).

displaced eastwards from the Gulf of Genoa, into the Gulf of Venice. Aegean Sea cyclones are principally winter and spring phenomena. In the Black Sea, cyclones form throughout the year, reaching a maximum of one per week in July and August in the eastern Black Sea.

Cyclogenesis Over North Africa During late winter and early spring, North Africa becomes the primary region of cyclogenesis in the Mediterranean basin. The cyclones are usually dry, but are characterized by high winds close to their centres. They move extremely rapidly, following an eastward track south of the Atlas Mountains before moving over the Mediterranean Sea across the coast of Tunisia at or near the Gulf of Gabes. Their speeds out of North Africa may be as high as 20–25 ms−1 . Lee cyclogenesis is thought to play an important role in the formation of these cyclones. The synoptic situation favouring development is the presence of an upper trough lying over Spain with its axis lying NE–SW, producing a deep south-westerly flow over north-west Africa. North African lows are associated with strong easterly to south-easterly winds over the southern Mediterranean and high seas in the Strait of Sicily. On occasions they may be responsible for the entrainment and transfer of

dust into and across the Mediterranean from the Sahara Desert, as illustrated in Chapter 14.

Other Regions of Cyclogenesis These include: r The formation of thermal low pressure centres over

the Iberian Peninsula in summer (HMSO 1962).

r Winter cyclogenesis can be expected to take place

over Cyprus when a cold front approaches the Anatolian Plateau from the north. r In the Middle East, over Syria and Iraq, cyclogenesis may occur in summer as an extension of the Indian monsoon. Despite this cyclone activity, the upper troposphere of the eastern Mediterranean is dominated by strong descent as part of the Asian monsoon system, and conditions remain dry (Trigo et al. 1999).

Frequency of Cyclone Formation Cyclone counts will vary depending upon the definition used and the method of detection. However, it is widely accepted that the most prolific centre of cyclogenesis is the Gulf of Genoa. The UK Meteorological Office (HMSO

76

Andrew Harding, Jean Palutikof, and Tom Holt

Fig. 3.4. (a) Regions of cyclone genesis and dominant cyclone tracks in the Mediterranean (modified from HMSO 1962) and (b) a TERRA satellite image of a cyclone centred on the Ionian Sea, 24 February 2006. , accessed 20 November 2008).

The Climate System

1962) estimated the number of cyclones forming in an average year at this location at 60, compared to 51 in the Ionian Sea, 28 in the Middle East, and 14 in the Sahara. The principal areas of cyclogenesis and the tracks of depressions are shown in Figure 3.4a and a large winter cyclone centred on the Ionian Sea to the south of Italy is shown in Figure 3.4b.

Local Winds The complex geography of the Mediterranean basin creates many local wind systems, due to the interplay of the atmospheric circulation described above with the distribution of land and sea, and of mountains and coastal plains in close proximity. These local wind systems include the Levanter and Vendavales in the Strait of Gibraltar, the Bora in the former Yugoslavia and the northern Adriatic, the Sirocco over North Africa, and the Khamsin over Egypt into the eastern Mediterranean. Table 3.2 lists the local wind systems of the Mediterranean region. Many of these local winds are associated with cyclogenesis and the passage of depressions, as described above. Some can only form in the presence of low pressure systems whilst others require these systems to achieve their maximum development. Some local winds bring welcome relief in the summer heat. The wind known as the Etesian in Greece and the Metemi in Turkey forms when the presence of a high pressure system over Hungary and a low pressure system over Turkey channels a cool northerly airflow over Greece and Turkey in the summer. Others are regarded with fear. The Khamsin flows out of the desert bringing desiccating dust-laden air in late spring and early summer which can obscure visibility and irritate the eyes, nose, and mouth. Many local winds are katabatic, i.e. they are a downflow of higher-density cold air. The Bora forms when cold air ponds over the mountainous regions of the former Yugoslavia, eventually spilling over the high mountain passes and flowing down into the northern Adriatic. It is most common in winter and wind speeds can exceed 30 ms−1 . The Bora wind can also affect Venice in winter, causing damage and disrupting water traffic. The most severe local winds occur when the mesoscale atmospheric circulation interacts favourably with the regional topography, and the Bora is no exception. In Croatia, the Bora wind of 14/15 November 2004 was intensified by the presence of a deepening low pressure system over the southern Adriatic, and a strengthening of the central European anticyclone.

77

It killed two people and injured over fifty. It disrupted energy supply and transport, and caused extensive damage to houses, harbours, and trees, with many olive trees uprooted. The Mistral is a katabatic cold wind of the western Mediterranean, most common in winter and spring. Cold air ponds over the Massif Central and Alps, and flows down the Garonne and Rhône valleys into the Gulf of Lion. Up to half of winter days in Marseilles can be affected by Mistral winds, which bring dry, cold, but sunny weather. The effects of the Mistral on the sea can sometimes be felt past Sicily, commonly causing wave heights of 5 to 6 m, and exceptionally up to 9 m. Although Mistral conditions can develop locally along the south coast of France, major Genoa cyclogenesis (see above) is necessary for an extensive Mistral to occur. Even after a Genoa low has moved eastwards, the Mistral can continue associated with a residual trough to the south of the Alps.

Extreme Events We have already described some severe wind events associated with local wind regimes in the Mediterranean. But, in people’s perception, the Mediterranean region is associated much more with other climaterelated hazards such as heatwaves and droughts, and with flash floods. Note that Mediterranean storms and floods are covered in greater detail in Chapter 18 in Part III of this volume on Natural Hazards.

Heatwaves Heatwaves in the Mediterranean have a number of impacts. They increase the number of deaths from heat stress, especially amongst the elderly, they increase energy demand for air conditioning, and they lead to widespread and devastating forest fires. Perhaps the most severe heatwave in the region was that of 2003, which affected the whole of Europe, and is considered to be the warmest summer since 1500 (Stott et al. 2004). The number of excess deaths in that heatwave around the Mediterranean is estimated at around 20,000 in Italy, 8,500 in Spain, and 2,000 in Greece. In Spain, mean daily temperatures of 33◦ C and over were recorded for at least half of the days (46/92 days) of the period June–August 2003 in 15 out of 48 cities. In 8 of these 15 cities, temperatures over 33◦ C were registered for more than 60 of the 92 days in the period (Simón et al. 2005). In Italy, daytime maximum temperatures remained between 38 and 40◦ C in most cities for weeks. As a result of the

78

Andrew Harding, Jean Palutikof, and Tom Holt TABLE 3.2. Local winds of the Mediterranean Location

Wind

France

Mistral

France

Bize

Spain

Levante

Spain

Leveche

Spain

Solano

Spain

Tramontana

Spain

Galerna

Spain

Cierzo

Italy

Maestro

Italy

Gregale

Balkans

Etesian

Balkans

Bora

Balkans

Varadarac

Africa

Khamsin

Africa

Scirocco

Africa

Ghibli

Characteristics • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • • •

Cold, dry, northerly or north-westerly katabatic flow Blows along the southern coast, as far as Genoa, most prevalent between Montpellier and Toulon Accompanied by clear weather and bright sunshine Associated with depressions in the Gulf of Genoa and a high pressure ridge from the Azores Cold, dry, northerly, north-easterly or north-westerly winter wind Blows in the mountainous regions of southern France in Languedoc Accompanied by heavy clouds Strong, moist, cool, north-easterly spring (Feb.–May) and autumn (Oct.–Dec.) wind Blows on the east and south coasts of Spain Associated with the Azores anticyclone Hot, dry, south-westerly wind Blows across south-east Spain Occurs in front of eastward moving depressions Hot, humid south-easterly or easterly summer wind Occurs on the east coast of Spain and in the Strait of Gibraltar Cold, dry, northerly; a form of Bora Blows along the northern Catalan coast and parts of Italy Cold, dry, north-westerly, all year but especially in winter Blows onto the north Spanish coast from the Atlantic Cold, dry, north-westerly, active for an extended (6-month) winter period Blows along the Ebro valley North-westerly summer wind Blows in the Adriatic when pressure is low over the Balkans Associated with fine weather and light clouds Strong, cold north-easterly wind (mainly winter) Blows in the Ionian Sea, usually lasting 2–3 days Frequently reaches gale force, usually accompanied by rain Northerly, summer, blows across the Aegean Associated with the Azores anticyclone and the Asian low Accompanied by high rates of evaporation and pleasant cool conditions Cold and dry, often gusty north-westerly winter katabatic wind Blows on the coast of Dalmatia into the Gulf of Trieste Associated with an eastward moving depression forced over the Balkans and cold air from Russia Similar to the Bora Blows along the Vardar valley into the Gulf of Thessaloniki Hot, dry, dusty, southerly Egyptian wind Associated with a trough moving across the eastern Mediterranean Accompanied by humidity drops to less than 10% Hot, very dry south to south-westerly wind that becomes moist as it traverses the Mediterranean Sea Comes from the desert climates of North Africa and the Near East, gathering moisture before reaching the Mediterranean • Blows along the majority of Mediterranean facing coasts, particularly of Algeria, Italy, and the Levant region • Very similar to the Scirocco but originating over Libya

Source: Harding (2006), modified from Rudloff (1981), Barry and Chorley (1992), and Kostopoulou (2003).

heatwave, and in the knowledge that such events are likely to become more common, many governments have put in place emergency plans to cope with future occurrences. It is likely that the 2003 heatwave is part of a rising trend, with a strong probability that this is related to global warming caused by human influence on the climate. Baldi et al. (2006) studied heatwaves between 1950 and 2000 in Italy; defining heatwaves as events that exceed the 90th percentile of the temperature

distribution (calculated over the period 1961–90) for six or more consecutive days. Table 3.3 shows the number of heatwaves per decade since 1950. It is clear that there has been a marked increase since 1990. Stott et al. (2004: 610) state that it is ‘very likely (confidence level >90%) that human influence has at least doubled the risk of a heatwave exceeding this threshold’. Half of all summers are expected to be as warm as 2003 by 2040, and the 2003 summer may approach the norm by 2080.

The Climate System TABLE 3.3. Decadal distribution of heatwave days Decade

Days (no.)

%

1951–60 1961–70 1971–80 1981–90 1991–2000

66 38 18 98 187

16 9 4 24 46

Total

407

100

Source: Modified from Baldi et al. 2006.

Droughts Although heatwaves and drought generally go hand in hand, the most severe droughts are measured in terms of their persistence. Thus in the summer of 2005 a severe drought affected the western Mediterranean region, but this was the culmination of dry conditions that began in the autumn of the previous year. In Spain, rainfall between November 2004 and March 2005, the traditional months of reservoir and groundwater recharge, was the lowest since 1947 (the first year of record). As reservoir stocks fell to 50 per cent of normal levels, rationing was imposed by regional governments in the east of Spain (García Herrera et al. 2007). Since rainfall is decreasing across much of the Mediterranean, it is likely that droughts are becoming more frequent. However, there is little evidence that this is the case. In part, this is because drought depends not only on the amount of rainfall, but also on the level of abstraction. Abstractions in the Mediterranean are increasing as water demand rises under pressure from competing uses for irrigation, tourism, industry, and domestic needs (Chapter 21).

Intense Rainfall Events Intense rainfall events are important because they cause flash floods and landslides, with damage to infrastructure and loss of life. In natural river catchments between about 25 to 2,500 km2 in area a flash flood could occur following a rainstorm of more than 200 mm in less than six hours. In urban areas flooding could occur in a built-up area of 1 to 100 km2 following even shorter rain storms of around 50 mm in less than one hour. Such intense rainfall amounts are generally produced by stationary meso-scale convective systems. Table 3.4 gives some examples of recent flash floods in the Mediterranean (Chapter 18). Over time, it appears that the contribution of intense rainfall events to the total precipitation input has increased in much of the Mediterranean. For example,

79

in Italy, daily rainfall amounts in excess of 128 mm contributed 4 per cent of total rainfall in the 1990s, compared to only 1 per cent in the 1950s. A similar increased contribution has occurred in Spain from rain events greater than 64 mm per day (Alpert et al. 2002).

Past Climates of the Mediterranean Long-term changes in climates of the Mediterranean are dealt with in several chapters elsewhere in this volume. In particular, Chapters 2 and 4 examine the long-term regional climate dynamics and external forcing factors during the Quaternary Period and earlier parts of the Cenozoic, and Chapter 9 looks in detail at Holocene climate and environmental change across the region. In order to set the context for the Mediterranean climates of the present and future, we look briefly at what has happened in the region over the last millennium and, in more detail, at trends in the last 50–100 years.

Mediterranean Climates of the Last Millennium The major climatic episodes of the last millennium in Europe are the Medieval Warm Period (MWP), the Little Ice Age (LIA), and Current Warm Period. Whereas the MWP and the LIA were due to natural variability, in the case of the LIA solar variability and volcanic activity, the Current Warm Period is widely considered to be due to human activities causing global warming (see below). The MWP lasted from the tenth to the fourteenth century. The LIA lasted from around the fourteenth to the end of the nineteenth century. Both the MWP and LIA are visible in the records of past climates from the Mediterranean. In north-west Spain, Martinez-Cortizas et al. (1999) found that the MWP was around 1.5◦ C warmer than the present day. In the eastern Mediterranean, it was a period of wetter conditions with, for example, high water levels in the Dead Sea and Sea of Galilee (Schilman et al. 2001). The LIA was a period of glacier advance in the Apennines and Pyrenees, and wintertime temperatures sometimes as much as 3◦ C colder than at present (Giraudi 2005). The impacts of the LIA on fluvial systems and glacial systems in the Mediterranean are presented in Chapters 11 and 12 respectively.

Climates of the Last Century We can trace the late-nineteenth-century warming that has taken place across the Mediterranean since the end

80

Andrew Harding, Jean Palutikof, and Tom Holt

of the LIA. Glaciers in the Pyrenees have shrunk substantially since that time. Total surface area has dropped from 40 km2 in 1895 to 30 km2 in 1958, 15 km2 in 1975, and just 8 km2 in 1992. In 1975 there were seventy glaciers in the Pyrenees but by 1992 this had shrunk to forty-one (Serrat and Ventura 1992). Today, the most southerly glacier in Europe is the Ghiacciaio del Calderone glacier, in the Apennines. Between 1884 and 1990, the area of this glacier shrank by half (D’Orefice et al. 2000) (Chapter 12). The Third Assessment of the Inter-governmental Panel on Climate Change stated that ‘the global average surface temperature has increased over the 20th century by about 0.6◦ C’ and that ‘there is new and stronger evidence that most of the warming observed over the last 50 years is attributable to human activities’ (IPCC Working Group I 2001). This warming is not uniform, nor is it constant over time. Using gridded temperature data, the IPCC found that warming has occurred over the Mediterranean in all seasons since the mid-1970s (Folland et al. 2001). However, more detailed analyses suggest more complex trends. Using station data for summer for the period 1950–99, Xoplaki (2002) found warming in the western and central Mediterranean, but cooling in the interior of eastern landmasses, especially eastern Turkey and the Balkans. Kutiel and Maheras (1998) studied data from 1873 to 1989 and found warming over the year as a whole, which was stronger in the western Mediterranean (around 0.4◦ C per century) and weaker in the eastern Mediterranean (only 0.2◦ C per century). When these data were analysed on a seasonal basis, a cooling trend was found in autumn in the eastern Mediterranean. Rainfall appears to have been largely decreasing across the Mediterranean over the last half century, although the patterns are complex. In Italy and Spain, this decrease reached 10–20 per cent over the 1951– 95 period (Piervitali et al. 1998; Romero et al. 1998). Winter rainfall amounts in the western and northern Mediterranean are related to the North Atlantic Oscillation (see above), with correlation coefficients of around –0.6 to –0.7 (Palutikof 2003). Between 1960 and 2000 the winter NAO followed a gradual rising trend from strongly negative values in the early 1960s to strongly positive values in the 1990s, and since then has generally remained positive. Since the Mediterranean receives most of its rainfall in the winter season, it is likely that the declining rainfall trend in the western and northern Mediterranean is related to the trend in the NAO. It is not possible to judge whether the unusual long-term trend in the NAO starting in the 1960s is in any way related to global warming. In contrast, some parts of the eastern

Mediterranean, particularly southern and central Israel, have shown an upward trend in rainfall over the last fifty years (Ben-Gai et al. 1994). Norrant and Douguédroit (2005) examined data from sixty-three rainfall stations across the Mediterranean to elucidate any trends between 1950 and 2000. They found a generally decreasing trend in rainfall, especially in the winter months, although not always statistically significant. Statistically significant declining trends in rainfall were found for: r r r r r r

March in the Atlantic region, October in Mediterranean Spain, December in the Gulf of Lion and Gulf of Genoa, January, winter, and the year in Greece, winter and the year in Italy, winter in the Near East.

A statistically significant increasing rainfall trend was found in April in the Gulfs of Genoa and Lion. Figure 3.5 shows composite temperature trends for the Mediterranean since 1960 for the year as a whole, and for winter and summer. These are calculated from eighty-four stations for the whole region, of which forty-one are in the western region, twenty-five in the centre and eighteen in the east (Harding 2006). These clearly show a warming trend in the west and central Mediterranean, which is present both seasonally and annually. This trend is absent in the eastern Mediterranean. Figure 3.6 shows composite rainfall trends for the same regions and time periods (ibid.). A declining trend is apparent in winter in all regions, whereas summer rainfall amounts are stationary. This confirms the findings of Norrant and Douguédroit (2005).

Future Climate Change Human influences are expected to lead to changes in climate and rising sea level during the twenty-first century and beyond. Industrial processes, the internal combustion engine, intensive agriculture, and deforestation all add greenhouse gases to the atmosphere, including carbon dioxide, the nitrous oxides, and methane. Whereas short-wave radiation from the sun is largely unaffected by these gases, long-wave radiation from the earth is absorbed. As atmospheric concentrations of these gases increase, so more long-wave terrestrial radiation is absorbed, leading to heating of the atmosphere and changes in pressure patterns and rainfall because the amount of energy in the atmosphere is increased.

Whole Mediterranean 4

Western Mediterranean 4

Annual

Central Mediterranean 4

Annual

Eastern Mediterranean 4

Annual

2

2

2

2

0

0

0

0

-2

-2

-2

-2

-4 1960 4

1970

1980

1990

-4 2000 1960 4

Winter (DJF)

1970

1980

1990

-4 2000 1960 4

Winter (DJF)

1970

1980

1990

-4 2000 1960 4

Winter (DJF)

2

2

2

2

0

0

0

0

-2

-2

-2

-2

-4 1960 4

1970

1980

1990

-4 2000 1960 4

Summer (JJA)

1970

1980

1990

-4 2000 1960 4

Summer (JJA)

1970

1980

1990

2000

4

Summer (JJA)

2

2

2

0

0

0

0

-2

-2

-2

-2

1980 Year

1990

-4 2000 1960

1970

1980 Year

1990

-4 2000 1960

1970

1980 Year

1990

2000

1980

1990

2000

1970

1980

1990

2000

1990

2000

Summer (JJA)

-4 1960

1970

1980 Year

Fig. 3.5. Composite graphs of annual, winter (DJF), and summer (JJA) temperature in the whole Mediterranean and the western, central, and eastern basins, 1960–2000. Units are standard deviations.

The Climate System

1970

1970

Winter (DJF)

-4 1960

2

-4 1960

Annual

81

82

Western Mediterranean 4

Annual

Central Mediterranean 4

Annual

Eastern Mediterranean 4

Annual

2

2

2

2

0

0

0

0

-2

-2

-2

-2

-4 1960 4

1970

1980

1990

2000

-4 1960 4

Winter (DJF)

1970

1980

1990

2000

-4 1960 4

Winter (DJF)

1970

1980

1990

-4 2000 1960 4

Winter (DJF)

2

2

2

2

0

0

0

0

-2

-2

-2

-2

-4 1960 4

1970

1980

1990

2000

-4 1960

1970

1980

1990

2000

1970

1980

1990

2000

1970

-4 1960

1970

Summer (JJA)

Summer (JJA)

Summer (JJA)

2

2

2

0

0

0

0

-2

-2

-2

-2

1970

1980 Year

1990

2000

-4 1960

1970

1980 Year

1990

2000

-4 1960

1990

2000

1980

1990

2000

1980 Year

1990

2000

Winter (DJF))

2

-4 1960

1980

4

4

4

Summer (JJA)

-4 1960

Annual

1970

1980 Year

1990

2000

-4 1960

1970

Fig. 3.6. Composite graphs of annual, winter (DJF), and summer (JJA) precipitation in the whole Mediterranean and the western, central, and eastern basins, 1960–2000. Units are standard deviations.

Andrew Harding, Jean Palutikof, and Tom Holt

Whole Mediterranean 4

The Climate System

The present-day climate of the Mediterranean already challenges economic activities in the region and puts ecosystems at risk. For example: r Rainfall barely provides sufficient water to support

many ecosystems and economic activities. Expensive irrigation systems and water transfer systems are required to maintain agricultural productivity in countries such as Spain. r Temperatures are high in summer, and any further warming could lead to heat stress, the need for air conditioning, and outdoor conditions too uncomfortable to sustain beach tourism. r Increases in damaging extremes (e.g. flash floods, droughts, and wildfires) could lead to further loss of life and economic damages. r Some coastal areas of the Mediterranean are already affected by flooding from the sea, particularly along the eastern coast of Italy. Venice experiences flooding due to a combination of subsidence and a rise in the levels of the lagoon waters. It is expected that the city will be uninhabitable at the end of this century unless new methods of protection from the water are installed. It is therefore very important to understand how future climates might change in the Mediterranean in response to global warming, and what the local impacts could be (Jeftic et al. 1992, 1996). The Inter-governmental Panel on Climate Change (IPCC) Third Assessment Report looked at future climate change at the regional scale. For the Mediterranean it found that: r The Mediterranean region is expected to warm at a

rate greater than global mean warming. Warming in the summer season (June to August) is expected to be as much as 40 per cent above the global mean. Taking into account the range of uncertainty, this implies a summer warming of at least 2◦ C, and possibly as much as 6◦ C by 2100. r Little change in winter-time rainfall is expected although summers are expected to become considerably drier (Giorgi et al. 2001). These two facts are not unrelated. Summer drying will mean lower evaporation rates. Evaporation requires latent heat; therefore energy which would be required to enable evaporation to take place will be available for warming the air. Summer warming is therefore predicted to be considerably above the global mean. Globally, sea level is likely to increase by anything between 30 cm and 50 cm by 2100, although the rise could be anything between 9 cm and 88 cm because of

83

the uncertainties involved in making the estimate (IPCC Working Group I 2001). The regional changes will be different, depending on local land movements (Chapter 13). However, sea-level rise is a threat to coastal regions of the Mediterranean including Venice. All areas of high population density and/or heavy infrastructure development are vulnerable, especially where the land is subsiding. Such a location is the Nile Delta, where many millions of people are involved in farming, and fishing in the coastal lagoons. Their livelihoods would be severely threatened by sea-level rise.

Local Changes in Mediterranean Climates So far we have looked only at Mediterranean-wide changes. However, we can examine regional variations using climate models; these are computer-based simulations of the behaviour of the atmosphere, oceans, land surface, and ice, and their interactions. By changing the concentrations of greenhouse gases in these models, future changes in climate can be simulated. To explore future climate changes in the Mediterranean we use output from the Hadley Centre Regional Climate Model, HadRM3 (Hudson and Jones 2002). It is important to note that these simulations are just one realization of a wide range of possible futures. Such a realization is generally called a scenario, or plausible future climate. Different assumptions about future economic activities, and hence future emissions of greenhouse gases and their concentrations in the atmosphere, can be made. The Special Report on Emissions Scenarios (Nakicenovic and Swart 2000) defined four future pathways of economic and demographic development up to 2100, with associated emissions scenarios. Here, we look at the A2 pathway, which defines a future world of very rapid economic growth, low population growth, and rapid introduction of new and more efficient technology. Major underlying themes are economic and cultural convergence and capacity-building, with a substantial reduction in regional differences in per capita income. Figure 3.7 shows plausible future changes in winter and summer mean temperature and rainfall across the Mediterranean in 2070–99 compared to 1961–90, based on the HadRM3 model using an A2 emissions scenario. Winter temperature increases of up to 5◦ C are simulated over land, with the largest increases over central Europe, Turkey, and the Middle East, and the lowest changes (2–3◦ C) over the maritime coastline of Europe. Summer temperature increases are larger, up to 7◦ C by

84

Andrew Harding, Jean Palutikof, and Tom Holt

50°N

5

50°N

60 40

4

20 3

40°N

40°N 0

2 1

30°N 10°W

0°

10°E

20°E

30°E

-20

10°W

40°E

50°N

-40

30°N

7

0°

10°E

20°E

30°E

40°E

50°N

0

6

-10

5 4

40°N

10°W

0°

10°E

20°E

30°E

-30

3

-40

2

-50

1

30°N

-20 40°N

40°E

-60

30°N 10°W

0°

10°E

20°E

30°E

40°E

Fig. 3.7. Future changes in temperature (left-hand side) and precipitation (right-hand side) over the Mediterranean (2070–99 minus 1961–90). Winter (DJF) changes are shown above, and summer (JJA) changes are shown below. Units are ◦ C for temperature and mm for precipitation. Changes are shown for the SRES A2 scenario in the Hadley Centre Regional Climate Model HadRM3—see text for further explanation.

2070–99, and greatest over the northern shore of the Mediterranean. Spatial variations of precipitation changes are, as expected, greater and less coherent than for temperature. In winter, the largest reductions in rainfall are seen over the Mediterranean coast and islands, extending inland to include much of Greece, Turkey, and southern and central Spain. Modelled reductions in average winter rainfall are generally of the order of −20 mm over the three month period in these areas, but can be up to −40 mm in isolated areas in southern Turkey and Greece. Increased rainfall is suggested over the northern shore of the Adriatic, into the Alps and central and northern France, with increases as great as +60 mm over the Alps. In summer, rainfall is expected to remain unchanged or decrease, with no areas of increased rainfall indicated. The greatest reductions are seen over the Alps, in a belt which extends southwest to include the Pyrenees, to as much as −60 mm over three summer months. Smaller declines are projected over North Africa and the Middle East.

Future Changes in Extremes In a region that can experience several rain-free months per year, and where intense rainfall events can lead to

devastating flash floods, any future changes in these extremes could be very damaging. Climate models such as HadRM3 can provide the necessary information to explore future changes in extreme events. Figure 3.8(a) shows the changes suggested by HadRM3 in the length of the summer drought in the Mediterranean. Most places around the Mediterranean experience a prolonged period of summer drought every year, with rainfall being an exceptional occurrence in August. Any tendency for this dry period to prolong will have profound implications for water resources in the region. Figure 3.8(a) suggests that, in future, the summer drought period will lengthen in most land areas of the Mediterranean south of 40◦ N, typically by around ten days. This tendency is shown to be particularly severe over the Middle East and Egypt, reaching an increase of twenty to thirty days. North of 40◦ N, most areas show no change or a decrease in the length of the summer drought, by as much as ten to twelve days. Figure 3.8(b) shows the future change projected by HadRM3 in the maximum amount of rainfall received in a single five-day period. The spatial patterns are extremely fragmented, but throughout Spain, the southern shore of the Mediterranean, and the Middle East, the maximum five-day rainfall is suggested to remain unchanged or decrease. The principal land areas

The Climate System

85

(a) 50°N

30 20 10

40°N

0 -10

30°N 10°W

(b)

0°

10°E

20°E

30°E

40°E

50°N 40

20

0

40°N

-20

-40

30°N 10°W

0°

10°E

20°E

30°E

40°E

Fig. 3.8. Future changes in (a) length of the summer drought, in days; (b) maximum five-day precipitation, in mm, over the Mediterranean (2070–99 minus 1961–90). Changes are shown for the SRES A2 scenario in the Hadley Centre Regional Climate Model HadRM3—see text for further explanation.

where large increases are suggested are central Italy, the former Yugoslavia, and Switzerland, where increases of 30–40 mm are simulated.

Conclusions We have looked in this chapter at all aspects of climate over the Mediterranean in the recent past, the present, and future and we have outlined the major controlling factors in relation to the global climate system. In one chapter it is not possible to treat every aspect of climate in great detail. Those who seek a more detailed

treatment are directed to Mediterranean Climate Variability (Lionelli et al. 2006) and to a special issue of Global and Planetary Change entitled Mediterranean: Trends, Variability and Change (Piero et al. 2008). Here, the goal has been to present the importance of the climate of the Mediterranean to all aspects of life in the region. This importance derives from, first, rainfall amounts which are barely sufficient and, second, the risks associated with weather extremes. Rainfall amounts, and the strong seasonality of rainfall, mean that competition for water is intense, and a limiting factor for regional development. In Catalonia,

86

Andrew Harding, Jean Palutikof, and Tom Holt TABLE 3.4. Summary of some recent flash floods in Mediterranean Europe

Place

Date

Barcelona area, Spain Nîmes, France

25/9/1962

Vaison-la-Romaine, France Brig, Switzerland

26/9/1992

Versilia, Italy

19/6/1996

Biescas, Spain

7/8/1996

Corbières, France

12–13/11/1999

Sovearto, Italy

9–10/9/2000

Anduze, France

8–9/9/2002

4–5/10/1988

22–4/9/1993

Event

Loss of life

250 mm of rainfall in 2 hours caused flash flooding in the Besos River basin Urban flood from the Cadereaux watersheds (less than 50 km2 ). Peak flow in the city of about 1,000 m3 /s Over 300 mm of rainfall in 24 hours; peak flow of about 1,000 m3 /s at Vaison-la-Romaine 40 mm of rainfall in 24 hours on 23rd and 65 mm on 24th and maxima at Simplon (in Saltina River catchment above Brig) of 120 mm on 23rd and 220 mm on 24th. Flooding in Brig up to 3 m for 12 hours 400 mm of rain in less than 6 hours with maximum rainfall intensity of about 88 mm in 30 mins Over 250 mm of rainfall in 6 hours. Peak flow of 400–600 m3 in small upstream catchment of the Aras River (area 18 km2 ) devastated camping site located at outlet Rural flood of the Aude River (4,840 square km), 30–50% of peak flow being produced by a 123-km2 watershed (Guame et al. 2004) 350 mm of rain in 24 hours with a peak at 185 mm in 6 hours, causing landslides Over 600 mm of rain in 24 hours at Anduze; peak flow of the Gard River of 5,000 to 7,000 m3 /s for a 1,400 km2 basin

More than 1,000 deaths 11 deaths

Insured costs US $80 million €610 million

58 deaths 2 deaths

US $33 million

26 deaths

US $33 million

87 deaths

35 deaths

US $3 million

16 deaths (12 from a campsite) 25 deaths (in the same river 35 people were killed in 1958)

€1.2 billion

Source: http://natural-hazards.jrc.ec.europa.eu/activities_flood_flashflood.html accessed 31st March 2009.

Spain, 4.5 million people are affected by chronic water shortages, and the authorities are pressing for the construction of a pipeline to divert water from the Rhône in France to Barcelona. At the other end of the Mediterranean, Turkey has been accused by Syria and Iraq of depriving them of much-needed water, as it continues to build a series of dams along the Euphrates and Tigris. Turkey also plans to sell water from the Manavgat River (which flows south from the Anatolian Plateau into the Mediterranean) across the Middle East (Gruen 2000). The regional climate is characterized by extremes, which cause economic damage costing many millions of Euros and, in the most severe cases, fatalities. Large inter-annual variability is linked to the occurrence of extreme events such as droughts. Intense rainfall events, lasting only a few hours, are also a feature of the region’s climates (Table 3.4). Against this background of marginal water resources and damaging extremes, any tendency for the Mediterranean climate to change in future, whether for the better or the worse, is of intense interest.

Acknowledgements Original research presented in this chapter was carried out as part of the project ‘Modelling the Impact of Climate Extremes’

funded by the European Union under contract EVK20CT20010018 and a University of East Anglia research studentship. Climate model data was supplied by the Climate Impacts LINK Project (DEFRA Contract EPG 1/1/154) on behalf of the Hadley Centre and UK Met Office. We also thank Jamie Woodward and the external reviewer for reviewing the original manuscript.

References Alonso, S., Portela, A., and Ramis, C. (1994), 1st considerations on the structure and development of the Iberian Thermal Low-Pressure System. Annales Geophysicae-Atmospheres, Hydrospheres and Space Sciences 12: 457–68. Alpert, P., Ben-gai, T., Baharad, A., Benjamini, Y., Yekutieli, D., Colacino, M., Diodato, L., Ramis, C., Homar, V., Romero, R., Michaelides, S., and Manes, A. (2002), The paradoxical increase of Mediterranean extreme daily rainfall in spite of decrease in total values. Geophysical Research Letters 29/1: doi: 10.1029/2001GL013554. Balafoutis, Ch. J. and Makrogiannis, T. J. (2001), Analysis of a heat wave phenomenon over Greece and its implications for tourism and recreation. Paper presented at the 1st International Workshop on Climate, Tourism and Recreation, Halkidiki. Baldi, M., Dalu, G., Maracchi, G., Pasqui, M., and Cesarone, F. (2006), Heat waves in the Mediterranean: a local feature or a larger scale effect? International Journal of Climatology 26: 1477–87.

The Climate System Barry, R. G. and Chorley, R. J. (1992), Atmosphere, Weather and Climate. Routledge, London. BBC (British Broadcasting Corporation) (n.d.). World Water Crisis. , accessed 17 April 2006. (n.d.). BBC Weather Country Guide to Egypt, , accessed 7 September 2007. Ben-Gai, T., Bitan, A., Manes, A., and Alpert, P. (1994), Longterm changes in annual rainfall patterns in southern Israel. Theoretical and Applied Climatology 49: 59–67. Bremere, I., Kennedy, M., Stikker, A., and Schippers, J. (2001), How water scarcity will affect the growth in the desalination market in the coming 25 years. Desalination 138: 7–5. Bromwich, D. H., Rogers, A. N., Kallberg, P., Cullather, R. I., White, J. W. C., and Kreutz, K. J. (2000), ECMWF analyses and reanalyses depiction of ENSO signal in Antarctic precipitation. Journal of Climate 13: 1406–20. Burak, S., Dogan, E., and Gazioglu, C. (2004), Impact of urbanization and tourism on coastal environment. Ocean and Coastal Management 47: 515–27. Buzzi, A. and Tibaldi, S. (1978), Cyclogenesis in the lee of the Alps: A case study. Quarterly Journal of the Royal Meteorological Society 104: 271–87. Corte-Real, J., Sorani, R., and Conte, M. (1998), Climate change, in P. Mairota, J. B. Thornes, and N. Geeson (eds.), Atlas of Mediterranean Environments in Europe: The Desertification context. John Wiley, Chichester, 34–6. D’Orefice, M., Pecci, M., Smiraglia, C., and Ventura, R. (2000), Retreat of Mediterranean glaciers since the Little Ice Age: Case study of Ghiacciaio del Calderone, central Apennines, Italy. Arctic, Antarctic, and Alpine Research 32: 197–201. Egger, J., Alpert, P., Tafferner, A., and Ziv, B. (1995), Numerical experiments on the genesis of Sharav cyclones: idealised simulations. Tellus Series A: Dynamic Meteorology and Oceanography 47: 162–74. Eshel, G. and Farrell, B. F. (2000), Mechanisms of Eastern Mediterranean rainfall variability. Journal of the Atmospheric Sciences 57: 3219–32. Cane, M. A., and Farrell, B. F. (2000), Forecasting Eastern Mediterranean drought. Monthly Weather Review 128: 3618–30. Flocas, H. J. and Karacostas, T. S. (1996), Cyclogenesis over the Aegean Sea: identification and synoptic categories. Meteorological Applications 3: 53–61. Folland, C. K., Karl, T. R., et al. (2001), Observed climate variability and change, in J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu (eds.), Climate Change 2001: The Scientific Basis. Contribution of Working Group 1 to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. García-Herrera, R., Paredes, D., Trigo, R. M., Trigo, I. F., Hernández, E., Barriopedro, D., and Mendes, M. A. (2007), The outstanding 2004/05 drought in the Iberian Peninsula: associated atmospheric circulation. Journal of Hydrometeorology 8: 483–98. Gaume, E., Livet, M., Desbordes, M., and Villeneuve, J. P. (2004), Hydrological analysis of the River Aude, France, flash flood on 12 and 13 November, 1999. Journal of Hydrology 286: 135–54. Giorgi, F., Hewitson, B., et al. (2001), Regional climate information—evaluation and projections, in J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and

87

D. Xiaosu (eds.), Climate Change 2001: The Scientific Basis. Contribution of Working Group 1 to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Giraudi, C. (2005), Middle to Late Holocene glacial variations, periglacial processes and alluvial sedimentation on the higher Apennine massifs (Italy). Quaternary Research 64: 176–84. Grenon, M. and Batisse, M. (1989), Futures for the Mediterranean Basin: The Blue Plan. Oxford University Press, Oxford. Gruen, G. E. (2000), Turkish waters: source of regional conflict or catalyst for peace? Water, Air and Soil Pollution 123: 565–79. Harding, A. E. (2006), Changes in Mediterranean climate extremes: Patterns, causes and impacts of change. Doctoral Dissertation Ph.D., Climatic Research Unit, University of East Anglia. HMSO, Meteorological Office (1962), Weather in the Mediterranean, i. General Meteorology. HMSO, London. Hudson, D. A. and Jones, R. G. (2002), Regional Climate Model Simulations of Present-Day and Future Climates of Southern Africa, Hadley Centre Technical Note 39. Hadley Centre, Met Office, Exeter. Hurrell, J. W. (1995), Decadal Trends in the North-Atlantic Oscillation—Regional Temperatures and Precipitation. Science 269: 676–9. IPCC Working Group I (2001), Summary for Policymakers, in J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, and D. Xiaosu (eds.), Climate Change 2001: The Scientific Basis. Contribution of Working Group 1 to the Third Assessment Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Jeftic, L., Milliman, J. D., and Sestini, G. (1992), Climate Change in the Mediterranean, Edward Arnold, London. Keskes, S., and Pernetta, J. C. (1996), Climatic Change and the Mediterranean, ii. Edward Arnold, London. Kadioglu, M., Tulunay, Y., and Borhan, Y. (1999), Variability of Turkish precipitation compared to El Nino events. Geophysical Research Letters 26/11: 1597–600. Köppen, W. (1936), Das geographische System der Klimate, in W. Köppen and R. Geiger (eds.), Handbuch der Klimatologie, iii Gebrüder Borntraeger, Berlin. Kostopoulou, E. (2003), The relationships between atmospheric circulation patterns and surface climatic elements in the eastern Mediterranean. Doctoral Dissertation Ph.D., Climatic Research Unit, University of East Anglia. Kutiel, H. and Benaroch, Y. (2002), North Sea–Caspian Pattern (NCP)—an upper level atmospheric teleconnection affecting the Eastern Mediterranean: Identification and definition. Theoretical and Applied Climatology 71: 17–28. and Maheras, P. (1998), Variations in the temperature regime across the Mediterranean during the last century and their relationship with circulation indices. Theoretical and Applied Climatology 61: 39–53. Turkes, M., and Paz, S. (2002), North Sea Caspian Pattern (NCP)—an upper level atmospheric teleconnection affecting the eastern Mediterranean—implications on the regional climate. Theoretical and Applied Climatology 72: 173–92. Lagouvardos, K., Kotroni, V., Dobricic, S., Nickovic, S., and Kallos, G. (1996), The storm of October 21–22, 1994, over Greece: Observations and model results. Journal of Geophysical Research-Atmospheres 101(D21), 26217–26.

88

Andrew Harding, Jean Palutikof, and Tom Holt

Lionelli, P., Malanotte-Rizzoli, P., and Boscolo, R. (2006), Mediterranean Climate Variability, Developments in Earth and Environmental Sciences 4. Elsevier, Amsterdam. Maheras, P. and Kutiel, H. (1999), Spatial and temporal variations in the temperature regime in the Mediterranean and their relationship with circulation during the last century. International Journal of Climatology 19: 745–64. Mahovic, N. S., Drvar, D., and Vrsnak, D. P. (2005), EUMeTrain: Cyclogenesis in the Mediterranean, , accessed 7 September 2006. Martinez-Cortizas, A., Pontevedra-Pombal, X., Garcia-Rodeja, E., Novoa-Muñoz, J. C., and Shotyk, W. (1999), Mercury in a Spanish peat bog: Archive of climate change and atmospheric metal deposition. Science 284: 939–42. Nakicenovic, N. and Swart, R. (2000), Emissions Scenarios. Special Report of the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Norrant, C. and Douguédroit, A. (2005), Monthly and daily precipitation trends in the Mediterranean (1950–2000). Theoretical and Applied Climatology 83: doi 10.1007/s00704-0050163-y. Palutikof, J. (2003), Analysis of Mediterranean Climate Data: Measured and Modelled, in H.-J. Bolle, Mediterranean Climate: Variability and Trends, Springer, Berlin, 125–32. Peñarrocha, D., Estrela, M. J., and Millán, M. (2002), Classification of daily rainfall patterns in a Mediterranean area with extreme intensity levels: the Valencia region. International Journal of Climatology 22: 677–95. Piero, L., Serge, P., and Xavier, R. (eds.) (2008), Mediterranean climate: trends, variability, and change. Global and Planetary Change 63/2–3: 87–282. Piervitali, E., Colacino, M., and Conte, M. (1998), Rainfall over the Central-Western Mediterranean basin in the period 1951–1995. Part I: Precipitation trends. Nuovo Cimento Della Societa Italiana Di Fisica C-Geophysics and Space Physics 21: 331–44. Conte, M., and Colacino, M. (1999), Rainfall over the Central-Western Mediterranean basin in the period 1951– 1995. Part II: Precipitation scenarios. Nuovo Cimento Della Societa Italiana Di Fisica C-Geophysics and Space Physics 22: 649–61. Price, C., Stone, L., Huppert, A., Rajagopalan, B., and Alpert, P. (1998), A possible link between El Nino and precipitation in Israel. Geophysical Research Letters 25: 3963–6.

Radinovic, D. (1987), Mediterranean cyclones and their influence on the weather and climate. PSMP Report Series 24. World Meteorological Organisation, Geneva. Raicich, F., Pinardi, N., and Navarra, A. (2003), Teleconnections between Indian monsoon and Sahel rainfall and the Mediterranean. International Journal of Climatology 23: 173–86. Rodo, X., Baert, E., and Comin, F. A. (1997), Variations in seasonal rainfall in southern Europe during the present century: Relationships with the North Atlantic Oscillation and the El Nino Southern Oscillation. Climate Dynamics 13: 275–84. Romero, R., Guijarro, J. A., Ramis, C., and Alonso, S. (1998), A 30-year (1964–1993) daily rainfall data base for the Spanish Mediterranean regions: First exploratory study. International Journal of Climatology 18: 541–60. Rudloff, W. (1981), World Climates. Stuttgart, Wissenschaftliche Verlagsgesellschaft. Schilman, B., Bar-Matthews, M., Almogi-Labin, A., and Luz, B. (2001), Global climate instability reflected by Eastern Mediterranean marine records during the late Holocene. Palaeogeography, Palaeoclimatology, Palaeoecology 176: 157–76. Serrat, D. and Ventura, J. (1992), Glaciers of the Pyrenees, Spain and France. US Geological Survey Professional Paper 1386-E (Glaciers of Europe). Simón, F., Lopez-Abente, G., Ballester, E., and Marti, F. (2005), Mortality in Spain during the heatwaves of summer 2003. Euro Surveillance 10: 156–61. Stott, P. A., Stone, D. A., and Allen, M. R. (2004), Human contribution to the European heatwave of 2003. Nature 432: 610–14. Trigo, I. F., Bigg, G. R., and Davies, T. D. (2002), A climatology of cyclogenesis mechanisms in the Mediterranean. Monthly Weather Review 130: 549–69. Davies, T. D., and Bigg, G. R. (1999), Objective climatology of cyclones in the Mediterranean Region. Journal of Climate 12: 1685–96. Xoplaki, E. (2002), Climate variability over the Mediterranean. Doctoral Dissertation Ph.D., Institute of Geography, University of Bern. Zalidis, G., Stamatiadis, S., Takavakoglou, V., Eskridge, K., and Misopolinos, N. (2002), Impacts of agricultural practices on soil and water quality in the Mediterranean region and proposed assessment methodology. Agricultural Ecosystems and Environment 88: 137–46.

This chapter should be cited as follows Harding, A. E., Palutikof, J., and Holt, T. (2009), The climate system, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 69–88.

4

Cenozoic Climate and Vegetation Change Chronis Tzedakis

Introduction This chapter traces the history of vegetation change in the Mediterranean area in response to climate variability over the last 65 million years (Myr 1 ), with particular emphasis on the most recent part of the record. Compared to other continental areas of the globe, the Mediterranean region is somewhat unusual in the abundance of palaeobotanical information (especially palynological) that is available. This is a function of its geological setting, which in some cases has led to the relatively undisturbed accumulation of thick sedimentary sequences in tectonic and volcanic basins. These sequences have provided an opportunity to develop long records of vegetation change, sometimes extending over hundreds of thousands of years. In the marine realm, sedimentary records from the Mediterranean Sea are not only providing palaeoceanographic information but also beginning to yield palynological information, which can be placed directly within a chronological and palaeoclimatic framework. However, it is only in the uppermost part of the geological column (i.e. in the Quaternary) that there are enough records to construct a continuous thread of vegetation changes and allow meaningful comparisons between sites to determine local differences and transregional similarities (e.g. Magri et al. 2004). Moreover, the majority of terrestrial records extending before the Holocene are located in southern Europe, 1

‘Ma’ and ‘ka’ are used here to denote million and thousand years before present, respectively, while ‘Myr’ and ‘kyr’ are reserved for durations. Unless otherwise stated, calendar years are used.

while the coverage of the Near East is low and of North Africa even lower (Figure 4.1). The information available for earlier periods anywhere in the Mediterranean is fragmentary at best, with large parts of the record not represented. This means that despite the relative wealth of information, the palaeobotanical record from the Mediterranean region remains very much incomplete, with significant temporal and geographical gaps. Thus, instead of providing a linear narrative of the last c.65 Myr, the approach followed here is to structure this review into separate sections, each representing different scales of environmental variability, and attempt to establish the general pattern of vegetation responses to it. These environmental regimes are here defined as (1) mega-scale (long-term climate trends), (2) macroscale (orbitally driven (Milankovitch) changes), (3) meso-scale (sub-orbital variability), and (4) micro-scale (interannual variability) (Table 4.1). The first category refers to a shifting climatic mean forced by global plate tectonics, leading to a gradual transformation of continental and oceanic palaeogeography and changes in atmospheric greenhouse gases (Zachos et al. 2001). Oscillating about this mean are higherfrequency climate changes generated by variations in the Earth’s orbital geometry that affect the seasonal and latitudinal distribution of incoming solar radiation (Milankovitch 1930; Hays et al. 1976). Superimposed on these trends and rhythms are abrupt climate changes lasting centuries to millennia, whose origin is still debated but whose effects are becoming increasingly better documented (e.g. Voelker 2002; McManus et al. 1999). Finally, the last category refers to annual

90

Chronis Tzedakis

Ljubljan a

Bouchet/Ribain s

Lagaccion e

La B ord e

V. di Castiglione Banyole s

T. Philippo n

Monticchi o

40°N

Ioannina

Sile s MD95-2042

Pa dul

ODP976

Xinias

Va n

Ko pais

San Rafae l Dar Fatma

MD95-2043

Ghab

MD84-627

Tigalmin e Tigalmamine

Ma’ale Efrayim

Sore q

30° 0°

20°

40°E

Fig. 4.1. Location of sites discussed in the text.

TABLE 4.1. Scales of environmental variability and vegetation responses Environmental regimes Mega-scale (106–7 yr)

Macro-scale (104–5 yr) Meso-scale (102–3 yr) Micro-scale (100–1 yr)

Vegetation responses

Tectonic transformation of continental and oceanic palaeogeography, long-term changes in atmospheric and oceanic circulation patterns and changes in atmospheric concentration of greenhouse gases. Long-periodic orbital variations Orbitally driven (Milankovitch) climate variability, ice sheet build-up and decay Millennial/centennial sub-orbital climate variability Interannual climate variability, fire, volcanic eruptions, pathogenic attacks

to decadal changes that, however, are not considered here because the vast majority of the palaeorecords do not have the necessary stratigraphic detail to resolve them. Moreover, this chapter does not discuss the (mainly Holocene) record of anthropogenic impact on vegetation (see Part II), and concentrates instead on examining the vegetation response to natural climate variability. In the following sections, the global and Mediterranean environmental backgrounds and the response of vegetation are examined under each of the first three scales. In the final section, an attempt is made to consider the effects of interactions between the scales, on the duration of forested periods, the amplitude of suborbital-scale changes, genetic divergence, and the origin

Evolutionary changes (adaptation, speciation, extinction). Appearance of new biomes

Glacial–interglacial vegetation cycles (ecosystem change, individualistic response, refugia, migration, extirpations, divergence) Population contraction/expansion, replacement Disturbance, population collapse

of the mediterranean 2 climate rhythm and broadleaved evergreen sclerophylls.

Mega-scale (106–7 yr) Global Changes: Climate Plate tectonics is the main driver of the climate shifts observed at this scale leading to the gradual and mostly 2 The use of the term ‘Mediterranean’ (with initial capital) is reserved for the actual geographical region. The word begins in lower case when denoting mediterranean-type climates or vegetation that are not necessarily exclusive to this geographical area.

Cenozoic Climate and Vegetation Change

unidirectional modification of the Earth’s boundary conditions (Zachos et al. 2001). Most prominent amongst the changes over the last 65 Myr is the widening of the North Atlantic, the opening of the Antarctic gateways (Tasmanian and Drake passages), the collision of the Indian and Asian plates and uplift of the Himalayas and Tibetan Plateau, the emergence of the Isthmus of Panama and the decline in atmospheric CO2 (see review by Zachos et al. 2001 and references therein; Pagani et al. 2005). These changes contributed to long-term global cooling and cryospheric development, which, however, did not proceed gradually and uniformly but was punctuated by a series of steps, representing rapid transitions into new climate states (Kennett 1995). In addition, long-period orbital variations, such as the 1.2 Myr cycle in obliquity and the 2.3 Myr cycle in eccentricity, modulating the amplitude of higherfrequency astronomical changes, could also be included in this category (Lourens and Hilgen 1997). Indeed, some of the transitions into new climate states are associated with changes in the long-period orbital variations, suggesting that the Earth System may be more susceptible to change during such nodal points (e.g. Lourens and Hilgen 1997; Wade and Pälike 2004) Figure 4.2 shows a compilation of benthic oxygen isotope (‰18 Obenthic ) records by Zachos et al. (2001), which provides a global view of Cenozoic climate change. The first part of the Cenozoic was characterized by ice-free conditions. Warmest conditions were reached during the Early Eocene, followed by a cooling trend during which deep-sea temperatures dropped from 12 to 4.5◦ C. A prominent and rapid shift in ‰18 O values at 34 Ma signals the onset of major ice accumulation in East Antarctica (Shackleton and Kennett 1975). Ice sheets estimated at c.50 per cent of that of the present-day ice sheet persisted in Antarctica until 26 Ma. Atmospheric CO2 reconstructions (Pagani et al. 2005) show values fluctuating between 1,000 and 1,500 ppmv during the Middle to Late Eocene, with an overall downward trend. By 34 Ma, CO2 concentrations had reached levels sufficiently low to have triggered the initiation of ice accumulation in Antarctica, while by the latest Oligocene, CO2 concentrations had reached pre-industrial levels of ∼290 ppmv. In general, the co-evolution of ‰18 O records and CO2 concentrations during the Eocene and Oligocene suggests a close coupling between global climate and the carbon cycle, while this association appears to weaken in the Neogene (Pagani et al. 1999a , 2005). From 26 Ma to 15 Ma, warmer conditions reduced the extent of Antarctic ice, with the Antarctic ice sheet becoming re-established after that. From 10 to 5 Ma temperatures continued to drop, leading

91

to the development of the marine-based West Antarctic ice sheet and ice accumulation in the Arctic. Somewhat warmer conditions prevailed in the Early Pliocene (5.33–3.2 Ma), before a trend towards the intensification of Northern Hemisphere (NH) glaciation (3.2– 2.8 Ma) (Shackleton et al. 1984, 1995a ). The onset of glaciation requires temperatures that are low enough for snow precipitation and reduced summer snow melt, and a sufficient moisture supply. The gradual shoaling of the Central American Seaway (CAS) and the establishment of the modern Atlantic-Pacific salinity contrast between 4.7 and 4.2 Ma has been considered a key event, because it led to the intensification of the North Atlantic Thermohaline Circulation (NATHC) and to increased moisture transport to high latitudes (Haug and Tiedemann 1998). Although initially this may have led to warmer conditions, the change in boundary conditions may have preconditioned the climate system towards the onset of NH glaciation (Haug et al. 2001). The final closure of the CAS around 2.7 Ma may have contributed to the crossing of a critical threshold in moisture supply and the onset of glacial conditions (Bartoli et al. 2005). In addition, Haug et al. (2005) proposed that ocean stratification in the subarctic Pacific Ocean at 2.7 Ma led to increased seasonality, with summer warmth extending into the autumn and thus providing the moisture source for snow accumulation in North America. Recent modelling experiments provide support for the notion that the closure of the CAS led to an intensification of the North Atlantic circulation and to increased precipitation over Greenland and North America, but the simulated changes in ice volume appear too small (Lunt et al. 2008). The implication is that the CAS closure does not appear to be the primary factor that caused the onset of NH glaciation and that a more likely candidate may be decreasing atmospheric CO2 concentrations (Lunt et al. 2008). This may have combined with the modulation of the amplitude of the 41kyr obliquity cycle by the longer 1.2-Myr cycle, which led to more extreme changes in tilt angle after 3.1 Ma and therefore the occurrence of cold summers in higher latitudes (Berger and Loutre 1991). According to Lourens and Hilgen (1997) the influence of the long-periodic variations in obliquity is observed during several key moments in Neogene climate history. More specifically, high-amplitude variations in tilt connected with the 1.2 Myr cycle appear to be associated with sea-level lowstands of the short-term eustatic cycles of Haq et al. (1987), implying that these cycles are glacio-eustatic (Lourens and Hilgen 1997). During the Oligocene, Wade and Pälike (2004) found that glacial events (Oi) were related to the ∼405-kyr

92

Chronis Tzedakis

0

Vegetation in the Mediterranean

Epoch Climate events Pleistocene NH glaciation Pliocene

Glacial-interglacial vegetation cycles Expansion of open vegetation

W. Antarctic ice sheet 10

Mixed temperate

Miocene Mid-Miocene climatic optimum 20

Age (Ma)

Late Oligocene warming 30

Broad leaved evergreen/deciduous

Oligocene Antarctic glaciation

Paratropical

40 Eocene Early Eocene climatic optimum

50

Tropical

60 Palaeocene

7

6

5

4

3 2 δ18Obenthic (‰)

1

0

-1

-2

Fig. 4.2. Global oxygen isotope record based on data from more than 40 DSDP and ODP sites. The curve shown is a five-point running mean. NH = Northern Hemisphere. Also shown are some global climate events and Mediterranean vegetation changes (modified from Zachos et al. 2001).

eccentricity cycle, and also to the 1.2-Myr obliquity amplitude modulation cycle. In the Miocene, glacial events (Mi) are correlated with the 1.2 Myr cycle. The Tortonian–Messinian boundary (∼7.2 Ma), which coincides with the onset of a significant sea-level drop and in the Mediterranean is characterized by a notable increase in cold planktonic foraminifera (left-coiling neogloboquadrinid species), is also associated with the 1.2 Myr cycle (Lourens and Hilgen 1997). In the PlioPleistocene, the third-order eustatic changes at around 2.8 and 1.7 Ma also appear to be related to the 1.2 Myr obliquity cycle. These events are marked in the Mediterranean marine record by the first occurrence of Neogloboquadrina atlantica (s) during MIS G10 (Lourens et al. 1996) and a strong increase of N. pachyderma (s) from MIS 64 at the Plio-Pleistocene boundary as defined in the Vrica section, southern Italy (Lourens et al. 2004). Curiously, however, whereas in the Pliocene and Pleistocene, periods of obliquity maxima, modulated

by the 1.2 Myr cycle correspond to enhanced glaciation (Lourens and Hilgen 1997), in the Miocene and Oligocene the opposite situation is observed, and low obliquity amplitude variations correspond to glacial events (Turco et al. 2001; Wade and Pälike 2004).

Global Changes: Vegetation The long-term Cenozoic trend from greenhouse to icehouse conditions had a profound influence on the history of marine and terrestrial biota. The reduction in both sea surface and air temperatures led to decreased evaporation and reduced moisture availability. In addition, cryospheric development and high-latitude cooling led to a sharpening of the meridional thermal gradient, which in turn resulted in increased wind strength, intensification of atmospheric and oceanic circulation, and near-shore cold water upwelling in certain areas

Cenozoic Climate and Vegetation Change

(Kennett 1995). Aridification was a major consequence of these changes. This was further accentuated in certain places by tectonic processes through mountain building, producing orographic rain-shadow effects, or blocking the passage of monsoonal winds. All these changes promoted the expansion of open, herbaceous vegetation. Indeed, the palaeobotanical record provides clear evidence of the emergence of plant families whose members are predominantly herbs and shrubs during the course of the Cenozoic (Singh 1988). Most of these families appeared in the Late Cretaceous/Palaeocene, Eocene, and Oligocene but did not expand until the Miocene or even later. Until the end of the Eocene the world appears to have been almost completely forested, with low and middle latitudes dominated by evergreen mega- and mesophyllous angiosperms with some gymnosperms (Bredenkamp et al. 2002). Grasses evolved sometime between 70 and 55 Ma (Kellogg 2001; Jacobs 2004), but remained restricted until the Miocene. In Africa, grasses are consistently recorded in the Early Miocene (23–16 Ma), but began to expand after 16 Ma, with widespread savanna becoming established by 8 Ma (Jacobs 2004). The history of mid-latitude grasslands (or steppes) is more uncertain, with evidence pointing to a Late Miocene appearance in Asia and North America. The first unequivocal evidence for tundra is from the Late Pliocene (Wolfe 1985 and reference therein). A major evolutionary step in the grasses was the appearance of the C4 photosynthetic pathway, which can be viewed as a CO2 -concentrating mechanism that maximizes rates of photosynthesis by eliminating the effects of photorespiration. This provides a competitive advantage over plants with the more primitive C3 pathway when the ratio of atmospheric CO2 to O2 is low (e.g. Ehleringer and Björkman 1977; Ehleringer et al. 1991). At present, the distribution of C4 grasses is strongly correlated with regions of high minimum temperatures, strong seasonal precipitation and a wet and warm growing season; C4 species of grasses today dominate the prairies of North America, the grasslands of Africa, and the llanos and cerrados of South America (e.g. Collatz et al. 1998; F. I. Woodward et al. 2004). The earliest unequivocal fossil record of a leaf that can be designated as C4 is dated at 12.5 Ma (Nambudiri et al. 1978). Stable carbon isotopic ratios have been used to distinguish C4 from C3 plants and these provide an earlier first appearance date around 15 Ma (Kingston et al. 1994). Molecular clock estimates place the origin of C4 much earlier between 25 and 32 Ma, while the phylogeny of the grass family suggests that the C4 photosynthetic pathway evolved independently several times in different subfamilies (Kellogg 2001). Between 8 and 6 Ma, a

93

major expansion in C4 grasses across low latitudes has been inferred on the basis of distinct changes in ‰13 C of fossil mammal tooth enamel and palaeosol carbonates, but the causes of this expansion remain unclear. Cerling et al. (1997) have suggested that it may be related to a decrease in atmospheric CO2 concentrations below a critical level. On the other hand, Pagani et al. (1999b), using alkenone-based CO2 reconstructions, showed that atmospheric CO2 concentrations had already stabilized at pre-industrial levels of ∼290 ppmv before the C4 expansion. Thus, CO2 changes could not have been the forcing mechanism behind this ecological shift, which according to Pagani et al. (ibid.) was driven by the enhanced low latitude aridity or changes in seasonal precipitation patterns. If that were true, however, then separate C4 expansions should have happened several times earlier at relatively more arid locations (Cerling et al. 1997). It is of course possible that local earlier expansions did take place but have not yet been documented. It is also possible that instead of their Miocene expansion, it was the initial Cenozoic evolution of the C4 photosynthetic pathway between 25 and 32 Ma that was driven by a drop in atmospheric CO2 levels. Indeed recent reconstructions of atmospheric CO2 , spanning the entire interval 5–45 Ma, show that concentrations fell from ∼1,500 to 1,000 ppmv around 35 Ma to preindustrial levels by 25 Ma (Pagani et al. 2005). The level of 500 ppmv below which the C4 photosynthetic pathway is favoured over C3 was first breached c.30 Ma and this confluence may suggest a causal relationship (Pagani et al. 2005), although tighter chronological control on the timing of the C4 origin is needed to test this idea further. The Late Miocene expansion of C4 grasslands, on the other hand, has recently been attributed to a fundamental increase in fire regimes as a result of the onset of marked seasonality characterized by a warm, moist growing season with high biomass production, followed by a dry season that would convert the biomass into highly combustible fuel (Keeley and Rundel 2005). An important point that emerges is that the establishment of a frequent fire regime was a key factor not only for the expansion, but also for the maintenance of C4 grasslands to the present day (Keeley and Rundel 2005). This is supported by modelling experiments, showing that in the absence of fire, areas that are today dominated by C4 grasslands and savannas have the climatic potential to form forests (Bond et al. 2005).

The Mediterranean World The Cenozoic environmental changes outlined above occurred against a background of continuous

94

Chronis Tzedakis

geodynamic evolution in the Mediterranean region. Jurassic geological reconstructions indicate the presence of a wedge-shaped equatorial ocean, named Tethys, separating northern from southern continents (Smith et al. 1992). The opening of the Atlantic Ocean after 170 Ma led to the collision of Africa and Eurasia and the gradual elimination of the Tethys Ocean. Today, none of the former Early Jurassic or older Tethys ocean floor remains, with the western Mediterranean underlain by Miocene and younger crust, while the eastern Mediterranean is floored by mid-Jurassic or younger crust (Livermore and Smith 1985). The connection to the Indian Ocean was maintained until the Middle Miocene; the timing of the final closure of this eastern gateway is not well-constrained, with different studies providing estimates ranging from 16 to 11 Ma (e.g. Dercourt et al. 1993; Yilmaz 1993; Jacobs et al. 1996). The collision between Africa and Eurasia gave rise to new mountain chains (Chapter 1) and also led to the separation of the Mediterranean from an eastern European epicontinental sea, the Paratethys (whose remnants today are the Black, Caspian, and Aral Seas). The separate evolution of the Paratethys started as early as the Late Rupelian (30–28 Ma) (Dercourt et al. 1993), but was completed by the Early Tortonian (∼11.6 Ma) (e.g. Sprovieri et al. 2003). This meant that the Paratethys no longer received marine waters, while the Mediterranean was deprived of the fresh waters of the eastern European rivers, which emptied into the Paratethys (Hsü et al. 1977). In the western Mediterranean, the connection to the Atlantic was maintained through two gateways, the Betic Corridor through southern Spain and the Riffian Corridor through northern Morocco (Krijgsman 2002). Arguably the most dramatic episode in the history of the Mediterranean Sea is the Messinian Salinity Crisis (MSC) some time between 6 and 5 Ma, which refers to its isolation from the Atlantic Ocean and its partial desiccation. During this time, large saline lakes replaced marine basins while some areas dried out completely, leading to the emplacement of huge thicknesses of evaporitic deposits (e.g. Hsü et al. 1977; McKenzie et al. 1999). The causes of this isolation (glacio-eustatic vs tectonic) have been much debated (Chapter 1), but the controversy appears to have been resolved by the development of an astronomically calibrated chronology based on tuning of sedimentary cycles seen across the Mediterranean to orbital changes (Krijgsman et al. 1999). This shows that the beginning and end of the MSC occurred at 5.96 Ma and 5.33 Ma, respectively, and that these changes were synchronous throughout the Mediterranean. Moreover, this precise

chronology allowed the placing of these events within the benthic ‰18 O framework (e.g. Shackleton 1995a). This showed that the onset of the MSC was not related to prominent glacio-eustatic sea-level falls of either glacial stages TG22 or TG20 (Figure 4.3), and pointed to a predominantly tectonic explanation, although orbitally driven long-term changes in sea level may have played a part (Krijgsman et al. 1999). The end of the MSC may have been associated with sea-level rise of interglacial stage TG5 (McKenzie et al. 1999), although Vidal et al. (2002) point out that this was not as prominent a glacio-eustatic event as that of TG9 (Figure 4.3) and suggest that the end of the MSC was also not directly controlled by sea-level changes. Be that as it may, the so-called ‘terminal flood’ of 5.33 Ma reestablished the connection to the Atlantic through the Strait of Gibraltar and marks the onset of the Pliocene. During the MSC, the substantially reduced water body would have led to reduced moisture availability and aridification. It has also been suggested that the absence of a Mediterranean salty outflow to the North Atlantic may have led to a weakening of deep water formation and promoted cooler conditions (Vidal et al. 2002). In biogeographical terms, the desiccation permitted animal migration across land masses, while the terminal flood led to the isolation of animals on Mediterranean islands and subsequent evolutionary changes (Azzaroli and Guazzone 1979/80). In summary, the present Mediterranean can be viewed ‘as the most recent of a series of “Mediterraneans” whose shape and area have evolved rapidly throughout Mesozoic and Cenozoic time’ (Livermore and Smith 1985: 86). The familiar configuration of the basin is a geologically recent development, with the isolation of the Mediterranean as an inland sea occurring in the Middle Miocene and the present outline of land and sea emerging around 5 Ma. With regard to the major changes in vegetation in the Mediterranean during the course of the Tertiary, a synopsis can be provided through global vegetation reconstructions and overviews by Wolfe (1985), Janis (1993), and Willis and McElwain (2002). In the Early Eocene climatic optimum (∼50 Ma), the precursors of the Mediterranean lands were covered by a tropical rainforest, which extended to 50◦ N (Figure 4.4a). These evergreen and laurel forests are considered to have been derived from the ‘Palaeotropical Geoflora’ (or ‘Tethys flora’) and represent an important component of Tertiary vegetation in Europe (Mai 1989). A ‘paratropical’ rainforest (sensu Wolfe 1985) extended between 50◦ and 60–5◦ N, and up to 70◦ N in coastal areas; this contained a mixture of tropical and temperate elements with mangrove

Cenozoic Climate and Vegetation Change

95

TG5

TG9

‰ 2

2.5

3

TG22

δ18Obenthic

TG20

3.5

MSC

Pliocene

Miocene

4 4.0

4.2

4.4

4.6

4.8

5.0

5.2

5.4

5.6

5.8

6.0

6.2

6.4

6.6

6.8

7.0

Age (Ma) Fig. 4.3. A section of the compiled oxygen isotope record of Figure 4.2 (Zachos et al. 2001) for the interval 4–7 Ma. The duration of the Messinian Salinity Crisis is indicated (see Chapter 1), as are some of the prominent Marine Isotopic Stages (e.g. TG5) for that period.

swamps along the coasts and is classified under the summer-wet biome (Willis and McElwain 2002). In the continental interiors, a broad-leaved evergreen forest occurred between 60–5◦ N and 70◦ N, and finally north of 70◦ N formed a polar vegetation dominated by broadleaved deciduous trees and deciduous conifers (Wolfe 1985). During the cooling trend of the second half of the Eocene (Figure 4.4b), tropical rainforest in the Mediterranean region was gradually replaced by paratropical rainforest; broad-leaved evergreen forest migrated south of 60◦ N, while to the north of that extended a diverse mixed coniferous forest (Wolfe 1985). Following the Oligocene initiation of the East Antarctic ice sheet and further global cooling, vegetation in the Mediterranean region became increasingly dominated by broad-leaved deciduous forests, which extended over large areas of the NH (Figure 4.4c). The majority of the deciduous component represents Arctotertiary species, which had already begun entering Europe in the Palaeocene/Eocene, but during the course of the Oligocene invaded all European forest communities, producing very rich mixedmesophytic forests (Mai 1989). Broad-leaved evergreen vegetation gradually retreated south of 30◦ N (Wolfe 1985; Janis 1992) and became increasingly temperate in character (Mai 1989). The warmer conditions from 26 Ma to 15 Ma, culminating in the Mid-Miocene climatic optimum, led to a northward expansion of broad-

leaved evergreen forest up to 45◦ along the coasts (Wolfe 1985), but the ensuing steady decline in temperatures led to the near complete extinction of European laurophyll species (Mai 1989) and the establishment of mixed temperate woodlands (Quade et al. 1994 and references therein). The decline in temperature, the elimination of the Tethys Ocean and mountain building all contributed to the aridification of Africa and the expansion of open vegetation (Singh 1988). With regard to the state of vegetation during the MSC, palynological studies from the Tyrrhenian basin show the presence of deserts and saline terrain vegetation interspersed with riparian or delta forests in the former abyssal plain and lowlands; a thermophilous mid-altitude zone with subtropical and temperate elements and a montane zone are reconstructed further inland (see review by BertolaniMarchetti 1985). Of particular significance are palynological results from the Ptolemais and Servia basins in northern Greece, spanning three phases: before (6.75– 6.7 Ma and 6.33–6.28 Ma), during (5.44–5.21 Ma), and after (4.36–4.15 Ma) the MSC (Kloosterboer-van Hoeve 2000). Because these results come from the same area they can be used to assess directly the character of these periods without any complications arising from differences in local factors when attempting long-distance comparisons. Vegetation in the pre-MSC phase was characterized by a mixture of deciduous and coniferous

(a) 120°W

150°W

180°

150°E

120°E

90°W

90°E

60°W

60°E

30°W

0°

30°E

(b)

Mixed Coniferous Broad-leaved Deciduous Broad-leaved Evergreen Paratropical Tropical

(c)

Fig. 4.4. Northern Hemisphere palaeogeography and global vegetation maps for selected time slices in the Tertiary: (a) Early Eocene (∼50 Ma), (b) Middle/Late Eocene (∼37 Ma), and (c) Late Oligocene (∼27 Ma). The dashed line within the Broad-leaved Deciduous forest in Figure 4.4c indicates the border between Mixed Northern Hardwood forest in the north and the more southern Broad-leaved Deciduous forest. Modified from Wolfe (1985), with additional information from Janis (1992).

Cenozoic Climate and Vegetation Change

trees with an almost complete absence of herbs. During the MSC, the record is strikingly different, showing strong cyclic (precessional) changes between periods of herbaceous dominance and coniferous dominance; transitional intervals show increases in deciduous and evergreen Quercus. In the post-MSC phase, the overall abundance of herbs decreased (but not entirely) and periods of evergreen and deciduous Quercus woodland alternated with pine-dominated woodland. The overall impression is of subdued cyclicity with moist forest conditions before the MSC, prominent aridity and seasonality during the MSC, and the establishment of more modern open woodland following the MSC, but with seasonality persisting. Considerably more information is available for the Pliocene, providing an insight into the regional differentiation around the Mediterranean Basin (the reader is referred to reviews by Axelrod 1973 and Suc et al. 1995a, b). For the Early Pliocene (Zanclean (5.33–3.6 Ma)) the vegetation reconstructions have been organized into the following geographical quadrants. The north-western region (from Catalonia to central Italy) was characterized by a strong subtropical element. Taxodium swamps were found in coastal plains, mid-altitude forest belts were occupied by Sequoia, Cathaya, Cedrus, and Tsuga, and higher-altitude forest belts by Abies and Picea (Suc et al. 1995a, b). The south-western region was dominated by herbaceous xerophytic vegetation (Asteraceae, Plantago), including some subdesertic elements. Mediterranean plants were regularly represented and some tropical plants were present in the southernmost parts (ibid.). In the south-central region, evidence from Calabria and Sicily shows a gradually decreasing subtropical component, and the steady presence of mediterranean elements and herbaceous vegetation (Bertoldi et al. 1989). The north-eastern region had temperate forests alternating with pine-dominated forests or herbaceous vegetation (Drivaliari 1993; Kloosterboer-van Hoeve 2000). Finally, in the south-eastern region the Nile area was dominated by open vegetation, which included desertic elements (Drivaliari 1993). In the coastal plain of Israel, a mediterranean vegetation was established along with Artemisia (Horowitz 1974). An intriguing pollen record of Pliocene age from the Plateau Tufa, Kurkur Oasis (west of the Aswan Dam) shows the presence of taxa of the Arctotertiary Geoflora (Alnus, Corylus, Salix, Ostrya, Betula, Tilia, Aesculus, Carpinus, Quercus, Platanus, Ulmus) along with mediterranean (e.g. Oleaceae, Rhamnaceae, Cistaceae, Pistacia, Celtis) and subtropical African (e.g. Podocarpus, Phoenix, Loranthoideae) elements; the record also shows an increase

97

in halophytic and xerophytic elements (van Campo et al. 1968). The Plateau Tufa record is reminiscent of another record of Early Pleistocene age from the Hoggar Mountains (1,830m a.s.l.) in the central Sahara where pollen of Tilia, Quercus, Alnus, Juglans, Pterocarya, Zelkova, Ulmus, Picea, Pinus, Ostrya, Corylus, and Fraxinus were found along with mediterranean and tropical elements (van Campo et al. 1964). Together the two records lend support to the notion that elements of the Arctotertiary Geoflora had penetrated into the central Sahara at the end of the Tertiary (Axelrod 1973). Pollen records from the early Late Pliocene (Piacenzian, 3.60–2.59 Ma) show a gradual reduction of the subtropical component and an increase in herbaceous vegetation. The initiation of major NH glaciation is marked in the Mediterranean region by the expansion of steppe (Suc 1984) and the start of glacial–interglacial vegetation cycles. After the PlioPleistocene boundary (1.81 Ma) there is a further increase in the abundance of herbaceous vegetation during glacial intervals in the Mediterranean (Combourieu Nebout and Vergnaud-Grazzini 1991). In parallel with the intensification of glaciation during the course of the Late Pliocene and Pleistocene, pollen records show the progressive disappearance of subtropical taxa from the Mediterranean and Europe. Several species disappeared right at the onset of major NH glaciation, while others persisted but at much reduced abundances. Over the next ∼2 Myr there was a continuous extinction of tree species from Europe. By about 600–400 ka (depending on location), the composition of European forests was similar to the present-day situation. The causes for these extirpations will be discussed in detail in the macro-scale section. In floristic terms, the present-day sclerophyll elements in the Mediterranean are considered to have their origins in the Tertiary laurophyll vegetation (Mai 1989). Palamarev (1989) provides a survey of the fossil occurrence of plant species of the ancient Tethyan and Paratethyan communities, which developed into the Mediterranean forest vegetation by the end of the Tertiary (Table 4.2). According to this, a small number of species appeared first in the Eocene, while twenty-three species have an Oligocene first appearance. The Miocene list of taxa is the most diverse with forty-nine new species, while five new species make their first appearance in the Pliocene fossil record. These records show that the major part of palaeomediterranean woody species appeared in the Miocene mixed evergreen and deciduous forests (Palamarev 1989). The Late Tertiary sclerophyllous

98

Chronis Tzedakis TABLE 4.2. List of genera of species’ first fossil appearance in the Tertiary record (from Palamarev 1989)

Eocene Oligocene Miocene

Pliocene

Pinus, Tetraclinis, Periploca, Platanus, Rhamnus, Chamaerops, Smilax Abies, Cupressus, Picea, Pinus, Acer, Arbutus, Carpinus, Celtis, Ceratonia, Cercis, Coriaria, Cotinus, Laurus, Myrtus, Olea, Ostrya, Pistacia, Punica, Quercus(evergreen), Viburnum, Smilax Abies, Cedrus, Juniperus, Pinus, Acer, Alnus, Anagyris, Buxus, Cerasus, Colutea, Daphne, Ficus, Fraxinus, Lonicera, Marsdenia, Nerium, Paliurus, Phillyrea, Pistacia, Prunus, Pyracantha, Quercus (including semi-evergreen and deciduous), Rhus, Styrax, Tamarix, Viburnum, Vitex, Ruscus Ephedra, Juniperus, Ilex, Jasminum, Prunus

forests also contained deciduous oaks, which appeared in the subtropical flora of southern Europe in the Early Miocene as montane elements and then descended to lower belts (Palamarev 1989). While the sclerophyllous floral elements appeared early in the Cenozoic, they only became widely established in the Mediterranean after the disappearance of the laurophyll vegetation (Mai 1989). In terms of the history of C4 grasses in the Mediterranean, stable isotopic evidence from palaeosol carbonates and fossil teeth in Greece show that Miocene and Pliocene vegetation was dominated by C3 plants, mainly representing forest and woodland with a small herbaceous component, with no evidence for the presence of C4 grasses (Quade et al. 1994). The present-day distribution of C4 plants in the Mediterranean is limited, presumably a reflection of the absence of a wet growing season: modern maps show minimal abundance in most places with somewhat elevated presence (<10% of ground cover) in north-west Africa and southern Iberia (e.g. F. I. Woodward et al. 2004).

Macro-scale (104–5 yr) Periodic variations in the Earth’s orbital parameters (eccentricity: ∼400- and 100-kyr periods; obliquity: 41 kyr; precession: 23 and 19 kyr) are the main forcing agent of the climate changes at this scale (see Chapter 2 for an in-depth discussion). While the Quaternary glacial–interglacial cycles represent an extreme expression of orbitally driven climate change, the effects of orbital variability are pervasive throughout the Cenozoic. For the ice-free part of the Cenozoic (pre-34 Ma), the available evidence suggests that climate change was dominated by variability in the precession bands (e.g. Cramer 2001). Obliquity is the primary beat of the glaciated part of the Cenozoic, underlining the sensitivity of ice sheets to high-latitude insolation changes associated with changes in the Earth’s tilt (e.g. Zachos et al. 2001). In addition, certain parts of the Oligocene and Miocene show a strong climate response that is lin-

early related to variations in the eccentricity band (both at 400- and 100-kyr periods) (e.g. Shackleton 2001; Wade and Pälike 2004). For the younger part of the Cenozoic, time-series analysis shows that Pliocene and Early Pleistocene ‰18 Obenthic records are dominated by variance in the obliquity band, while there is no wellmarked concentration of variance in the precession band (Shackleton et al. 1995a ). The Pliocene initiation of major NH glaciation (Shackleton et al. 1984; Raymo et al. 1989) marks the onset of 41-kyr glacial cycles, while in the Middle and Late Pleistocene ice volume variations become dominated by the 100-kyr cycle along with a precession component (Imbrie et al. 1984; Shackleton et al. 1990). In the Mediterranean, obliquity-related variations in planktonic ‰18 O records are recorded intermittently over the past 5.3 Ma (Lourens and Hilgen 1997) and reflect changes in global ice volume, but the signal is amplified by changes in regional salinity and/or temperature (Hilgen et al. 1993; Lourens et al. 1996). Obliquityrelated changes in sea surface temperature (SST) occur in phase with ‰18 O changes, suggesting that the glacial cycles influenced conditions in the Mediterranean Sea (Lourens et al. 1992). However, whereas equatorial Atlantic (e.g. Tiedemann et al. 1994) and Pacific ‰18 O records (e.g. Shackleton et al. 1990) show the influence of precessional cyclicity only in the last million years, planktonic ‰18 O and SST records from the Mediterranean are strongly influenced by precession over at least the last 12 million years (Turco et al. 2001), suggesting that the open ocean and Mediterranean systems operate partly independently (Lourens et al. 1992; Lourens and Hilgen 1997). The Mediterranean precessional component is linked to lithology of marine cores, with ‰18 O minima coinciding with sapropel layers and reflecting increases in freshwater input (e.g. Rossignol-Strick 1983; Lourens et al. 1992). This input is associated with periods of intensification of the African summer monsoon during precession minima, leading to light oxygen isotope anomalies and reduced deep water ventilation in the Mediterranean (Chapter 2). During

Cenozoic Climate and Vegetation Change

boreal summer, the northerly migration of the intertropical convergence zone (ITCZ) leads to increased heating of the North African land surface centred around 20◦ N and the drawing of maritime air from the Atlantic (deMenocal 2004). At times of minima in the precession index, boreal summer insolation maxima produce an enhanced land-ocean pressure difference and thus an intensified low-level monsoonal flow (Tuenter et al. 2003). This leads to the contraction of the desert belt (e.g. Roberts and Wright 1993), recharging of palaeolakes and reactivation of wadi systems in North Africa and the Levant (Chapter 2). The freshwater input into the Mediterranean associated with sapropel formation has been traditionally attributed to increased Nile discharge (e.g. Rossignol-Strick et al. 1982; Krom et al. 2002), but recent studies have suggested that northward penetration of the African monsoon beyond the central Saharan watershed at 21◦ N led to parallel discharge along the wider North African margin (Rohling et al. 2002a; Scrivner et al. 2004). The expansion of savanna-type vegetation over the Sahara during precession minima would also lead to increased cohesiveness of soil particles and reduced dust export to the Mediterranean (Middleton 1985 and see Chapter 14). Conversely, during precession maxima increased aridity and reduced vegetation cover in North Africa would provide increased dust supply to the Mediterranean (Chapter 14). Indeed records of dust deposition of the last 3 million years from marine cores in the Mediterranean contain a pervasive precessional signal (Lourens et al. 2001). The precessional changes in dust supply are modulated by the 100-kyr and 400-kyr eccentricity cycles, while a 41-kyr signal is also observed (prior to 0.95 Ma) as well as the 1.2-Myr modulation of obliquity (Larrasoaña et al. 2003). Similar orbital effects are observed in the character of sapropels, with small- and large-scale bundles of sapropels corresponding to maxima in the 100-kyr and 400-kyr eccentricity cycles (Hilgen 1991). In addition, sapropel thickness appears to be related to obliquity, with thick (thin) sapropels corresponding to obliquity maxima (minima) (Lourens et al. 1996). This influence of obliquity on sapropel formation is considered to proceed directly via summer insolation, rather than via glacial cycles (Lourens et al. 1996).

Sources of Confusion and Unresolved Issues The wealth of palaeoenvironmental information that has accumulated from the Mediterranean has undoubtedly contributed to an improved understanding of the complexity of climatic mechanisms that have operated

99

in a region located on the transition between high- and low-latitude influences (see Rohling et al. Chapter 3). However, in spite of (or even because of) this plethora of palaeodata, a number of confusing or unresolved issues remain. 1. The African monsoon and the Mediterranean. This is related to the notion that the African monsoon extended at certain times over the Mediterranean and has been responsible for the increase in precipitation during interglacials. The confusion arises from the close coupling of precession minima, boreal summer insolation maxima, northward migration of the ITCZ, intensification of the African monsoon, increased freshwater input into the Mediterranean Sea, reduced deep water ventilation and sapropel deposition, leading to the notion of a direct influence of the African monsoon on Mediterranean climate. In fact, the available evidence suggests that the connection is only indirect (i.e. increased runoff from the Nile and North Africa) and does not imply a northward extension of the monsoon over the Mediterranean. As discussed earlier, Rohling et al. (2002a ) have suggested that African monsoonal summer rains penetrated sufficiently far to the north during the last interglacial to cross the central Saharan watershed at ∼21◦ N and therefore some runoff would have reached the Mediterranean via northward-draining catchments, while archaeological evidence suggests that during the early Holocene monsoonal rains penetrated at least as far north as 25◦ N (Mandell and Simmons 2001). Model results, using extreme values of the last million years in precession and obliquity, show monsoonal precipitation penetrating beyond 25◦ N but still not reaching the North African coast (Tuenter et al. 2003). Indeed, BarMatthews (2000) have shown that the isotopic composition of precipitation in Israel during periods of late Quaternary sapropel deposition had a Mediterranean origin and was distinct from African monsoonal composition. In addition, work from the early Holocene in the northern Red Sea shows a humid interval corresponding to the deposition of sapropel S1 in the Mediterranean, but no clear evidence of surface water freshening in the central Red Sea, suggesting that the increased precipitation could not have originated from a northward extension of the monsoonal influence, but from the Mediterranean (Arz et al. 2003). 2. Precession minima, sapropel deposition and seasonality of precipitation in the Mediterranean during interglacials. A closely related issue to the environmental conditions surrounding sapropel deposition during interglacials is the amount and timing of precipitation in the Mediterranean. Isotopic evidence from Israel suggests a distinct increase in rainfall during precession minima coincident

100

Chronis Tzedakis

with sapropel deposition (Bar-Matthews 2000; Vaks et al. 2003). Indeed, an enhanced regional precipitation regime has been invoked in order to explain homogeneous sea surface salinities (SSS) across the Mediterranean (if freshwater input had originated mainly from Nile discharge then strong SSS gradients would have been observed) (Kallel et al. 1997). Moreover, this regime involves (1) not only a southern, but also a northern Mediterranean borderlands contribution (Rohling and Hilgen 1991) and (2) not only increased winter precipitation, but also a rise in summer rainfall with high moisture availability persisting throughout the year (Rossignol-Strick and Paterne 1999). The operation of a ‘Mediterranean monsoon’ (i.e. distinct from the African monsoon) as a result of elevated land–sea temperature contrast during summer insolation maxima in the south-east Mediterranean has been invoked in order to account for the early Holocene humid interval in the northern Red Sea (Arz et al. 2003). The concept of the disappearance of the characteristic mediterranean aridity during interglacial summer insolation maxima is thus being increasingly encountered in the literature, but this is inversely proportional to the evidence available to support it. The main argument for the increased summer rainfall appears to rest on palynological grounds from sapropel layers in marine cores showing high abundance of deciduous oak, which is considered to reflect annual precipitation levels on the order of 800–1,200 mm, providing year-round moisture (Rossignol-Strick 1999; Rossignol-Strick and Paterne 1999). There are at least three problems with this reasoning: (1) given that the species of deciduous oak cannot be palynologically resolved, the high oak values may reflect more drought-resistant species (even below 600 mm yr−1 ); (2) even if correct, values of 800– 1,200 mm yr−1 are encountered today in many places in the Mediterranean, but precipitation is still concentrated in the winter half of the year; and (3) equally high deciduous oak values are encountered at later times during the course of interglacials, which in turn begs the question: are summer-wet intervals a recurrent phenomenon within different parts of an interglacial? If this is the case, then different climatic mechanisms are required to explain this recurrence at different parts of the insolation cycle. In fact, closer examination of detailed terrestrial pollen sequences across the northern Mediterranean reveals that the timing of interglacial sapropel deposition corresponds to a characteristic peak in mediterranean sclerophylls (Olea, Pistacia, 3 3 In the increased-rainfall early interglacial scenario, a common assumption has been that the peak in Pistacia

Phillyrea, evergreen Quercus) (Rossignol-Strick 1995, 1999; Magri and Tzedakis 2000). More specifically, the highest sclerophyll abundances are encountered during the first part of the period of sapropel deposition (i.e. the first ∼2 kyr), before the expansion of more temperate vegetation (e.g. Ostrya, Carpinus). 4 This not only suggests the occurrence of relatively mild winters, but also firmly points to the presence of summer aridity, more extreme than today and perhaps extending over a larger part of the year. The presence of a Pistacia peak in the early Holocene was noted by RossignolStrick (1999), but was explained as indicating summer drought in the lowlands, while higher elevations received convective rain. The problem with this interpretation, however, is that it becomes difficult to argue for a major change in the character of Mediterranean climate affecting only higher elevations. In fact, the pollen record from the Ioannina basin, an intramontane plateau 470 m a.s.l., in north-west Greece, and with a pollen catchment including surrounding mountains rising to 1,800 m a.s.l., shows a distinct peak in mediterranean sclerophyll elements during the early part of the last interglacial coeval with the early phase of the sapropel S5 deposition (Figure 4.5) (Tzedakis et al. 2002b, 2003b). Parallel oxygen isotopic analyses from authigenic carbonate material (reflecting mainly a summer signal) from the same levels analysed for pollen, suggest a decrease in the precipitation/evaporation ratio during the pollen zone of mediterranean sclerophylls (Frogley et al. 1999; Tzedakis et al. 2003b). Moreover, a striking feature of early Holocene records from the drier parts of the Near East is that the expansion of woodland is often delayed for several millennia (e.g. van Zeist and Bottema 1991; Wick et al. 2003). This suggests that the increased and extended aridity during the boreal summer insolation maximum kept moisture levels below the tolerance threshold for tree growth, often until 8 ka, in the more ecologically stressed areas. In places where moisture availability was not limiting, such as southern Europe, woodland expansion occurred near the Pleistocene/Holocene boundary of 11.5 ka, with increased summer aridity leading to the expansion of mediterranean sclerophylls ∼10 ka. pollen represents an increase in populations of the deciduous type (P . terebinthus). However, the concomitant increase in Olea, Phillyrea, and evergreen Quercus would suggest the involvement of the evergreen species (P . lentiscus). 4 ‘Ostrya’ is here used for the Carpinus orientalis/Ostrya carpinifolia pollen type, ‘Carpinus’ for Carpinus betulus, and ‘chenopods’ for Chenopodiaceae/Amaranthaceae.

Cenozoic Climate and Vegetation Change Wm

101

-2

550

Insolation 65°N June

500 450 %

Temperate phase of succession

400

20 Abies 10

% 50

0 Carpinus betulus

40 30 20 10

% 20

0 Sclerophylls

10 % 80

0 Deciduous Quercus

60 40 20

% 100

0 AP-PJ

50

0

0 Late Interglacial 115

120

125

130

Age (ka)

Thus the palynological evidence suggests that, although the early part of interglacials was wetter than the preceding late glacial periods, this does not automatically translate to conditions of higher-thanpresent annual precipitation and a summer-wet regime. An apparent incongruity is therefore beginning to emerge for the early part of interglacials coeval with boreal summer insolation maxima, with isotopic evidence from speleothems indicating enhanced rainfall (Bar-Matthews 2000; Vaks et al. 2003), while the palynological evidence suggests extensive summer aridity. One way to reconcile the two data

Fig. 4.5. Summary pollen diagram showing the main taxa of the interglacial succession during the Late Interglacial at Ioannina (Tzedakis et al. 2002b, 2003b). AP-PJ is the arboreal pollen excluding Pinus and Juniperus (i.e. the ‘temperate’ taxa). Note different scales for each taxon. Also shown is the June insolation for 65◦ N (Berger 1978) and the duration of the temperate phase of succession (corresponding to the earlyand late-temperate zones of the Turner and West (1968) scheme).

sets is to invoke increased precipitation during the autumn/winter months. This would be a function of higher summer insolation regimes contributing to high sea surface temperatures persisting into the autumn and leading to higher levels of precipitation. Indeed, the increase in SSTs reconstructed during times of sapropel deposition has been considered to represent a summer signal under high summer insolation regimes (Lourens et al. 1992). An intriguing possibility is that this excess precipitation may have taken the form of severe storm events (Vaks et al. 2003). This is indicated by the large amount of detrital material incorporated in the Soreq

102

Chronis Tzedakis

cave speleothems and sharp decreases in concentrations of Sr, Ba, and U and in the ratios of 234 U/238 U and 87 Sr/86 Sr, which may reflect enhanced weathering of the host rock (Ayalon et al. 1999). Speculating further, high-magnitude storm events in the autumn season may have added to a freshening of the Mediterranean, but not contributed substantially to soil moisture availability for plant growth as most of the water would have been rapidly removed as fluvial runoff. Syntheses of lake-level records from the northern Mediterranean areas suggest wetter conditions during the early Holocene (Harrison et al. 1996). For reasons that are not entirely clear, since lake-level reconstructions from unlaminated sediment cores do not usually provide evidence on the seasonality of precipitation, these conditions have been attributed to increased summer precipitation, representing a local summer monsoon situation affecting the Iberian and Balkan peninsulas, but not Italy where conditions appear to have been drier (Harrison et al. 1996). However, Digerfeldt et al. (2000) have expressed concern over the reliability of lake-level reconstructions based on the lithology of a single sediment core from each basin and raised doubts over the accuracy of the Mediterranean lakelevel data, where few lakes are available and results often appear heterogeneous. In what is arguably the best study of lake-level reconstructions in the northern Mediterranean sector, Digerfeldt and colleagues (2000), using the technique that bears his name, undertook chemical and mineral analyses on multiple sediment cores which were pollen-stratigraphically correlated from the Xinias basin, central Greece. The results show a lake-level lowering towards the end of the Late Glacial (possibly coeval with the Younger Dryas chron), followed by a general rise in the early Holocene. This rise culminates around 8 ka, at the time of expansion of Ostrya populations (Digerfeldt et al. 2000; Digerfeldt, pers. comm.). This scenario is echoed in a combined isotopic, geochemical, palynological, and charcoal study from laminated sediments in Lake Van, south-east Turkey where the early Holocene is characterized by a gradual increase in moisture availability relative to the aridity of the Younger Dryas, but the wettest conditions were established after 8.2 ka (Wick et al. 2003). Thus, the available evidence suggests that at times of boreal summer insolation maxima in the early part of interglacial periods, northern Africa south of ∼24◦ N would have been under the influence of an enhanced summer monsoon (Kuper and Kröpelin 2006), while at least the northern Mediterranean sector would have experienced increased summer aridity. If an increase in precipitation (in any shape or form) did occur, that must

have been confined to the autumn/winter. The gradual rise in moisture levels recorded during the early part of interglacials suggests that the freshwater contribution from northern Mediterranean catchments could not have been a significant factor for the onset of sapropel deposition. This view appears to be in conflict with the ‰18 O record of the Argentarola Cave stalagmite on the Tyrrhenian coast of Italy, indicating a distinct increase in rainfall during Marine Isotopic Event 6.5 (∼175 ka) coeval with the deposition of sapropel S6 (Bard et al. 2002). However, S6 is an unusual ‘cold’ sapropel, occurring during a glacial stage and as such it may not necessarily be representative of precession minima associated with interglacial conditions. Clearly, additional speleothem and lake-level records from interglacials are needed to clarify the status of the northern Mediterranean freshwater contribution. 3. Mediterranean pluvials. The notion of glacial pluvial conditions, resulting from the deflection of moist westerlies southwards by the presence of ice sheets, has long been a part of the Mediterranean palaeoenvironmental narrative (see Butzer 1957 for an insightful early review). Arguably, some of the most convincing evidence for the presence of pluvial conditions is associated with the presence of extensive palaeolake systems along the Dead Sea Transform during the last glacial (e.g. Neev and Emery 1967; Begin et al. 1974). Supporting evidence was provided by syntheses of lake-level records from the northern Mediterranean sector, suggesting that during the Last Glacial Maximum (LGM 5 ) conditions were wetter than present (Harrison and Digerfeldt 1993; Harrison et al. 1996). By extension, this created an apparent incongruity when juxtaposed with pollen records from the Mediterranean which showed that landscapes were dominated by open vegetation (e.g. Elenga et al. 2000), indicating arid and cold conditions during this period. Over the years different climatic scenarios have been invoked to explain this apparent discrepancy between geomorphological and palaeoecological records, most notably (1) increased cloudiness and reduced evaporation rates coupled with lowered winter temperatures and precipitation levels (COHMAP Members 1988), or (2) cold winters, intense winter 5 The LGM, as defined by the EPILOG project, is the interval between 19,000 and 23,000 cal. yr BP (i.e. 16,100–19,500 14 C yr BP), centred on 21,000 cal. yr BP (18,000 14 C yr BP) (Mix et al. 2001). This interval is coeval with the lowest sea-level stand (120–5m below present (e.g. Fairbanks 1989)) of MIS 2 and is characterized by low temperatures and an absence of extensive millennial-scale climate variability (Mix et al. 2001) (Figure 4.6).

Cenozoic Climate and Vegetation Change

precipitation, and summer drought (Prentice et al. 1992). However, as discussed earlier, doubts have been raised over the reliability of lake-level reconstructions based on single sediment cores and the accuracy of data in the Mediterranean area (e.g. Digerfeldt et al. 2000) and have led to a re-evaluation of some of the key sites used in the LGM lake-level databases. A prominent example has been the re-examination of the Xinias basin, central Greece, which showed that lake levels at Xinias were low during the LGM (Digerfeldt et al. 2000). Another site central to the northern Mediterranean lake-level syntheses was the Ioannina basin, north-west Greece, where a series of lake margin or beach deposits from the Kastritsa cave, indicating higher lake levels than present, had been dated to the LGM (Higgs et al. 1967). While recent tectonic uplift of the area cannot be excluded (e.g. King and Bailey 1985; Tzedakis 1994), new AMS dates from the Kastritsa cave have nonetheless shown that the beach deposits are in fact older than the LGM (i.e. before 21 ± 2 ka) (Galanidou et al. 2000; Galanidou and Tzedakis 2001). Turning to the speleothem records from Israel, (BarMatthews et al. 1999) have inferred cold (∼12◦ C) and dry (∼250 mm yr−1 ) LGM conditions on the basis of the Soreq Cave isotopic record. Particularly enlightening has been the comparison between the speleothem records from Soreq Cave, located on the western side of the central ridge of Israel, which today receives large amounts of orographic rain originating from the Mediterranean, and the Ma’ale Efrayim Cave in the rain shadow to the east of the central ridge (Vaks et al. 2003). Today, a high evaporation to precipitation ratio leads to a negative water balance and thus very little water reaches the unsaturated zone in the Ma’ale Efrayim region, where no speleothem deposition occurs. Indeed, the record shows that no speleothem deposition took place in the Ma’ale Efrayim Cave either during the Holocene or MIS 5. By comparison, speleothem growth in this cave occurred during most of the glacial intervals of MIS 2, 3, 4, and 6, when temperatures and evaporation are considered to have been lower and effective moisture increased, allowing water to reside in the unsaturated zone (Vaks et al. 2003). However, within MIS 2, a hiatus in the record occurs between 24 and 19 ka, indicating a sharp drop in precipitation below the threshold required for speleothem deposition, and in agreement with the trends in the Soreq isotopic record to the west (Bar-Matthews et al. 1997). Comparison of the two cave records over the last 60 kyr shows similar climatic trends on both sides of the ridge and indicates that the origin of the precipitation was the eastern Mediterranean Sea (Vaks et al. 2003).

103

Faunal (planktonic foraminifera)-based reconstructions of LGM sea surface temperatures from the Mediterranean show the largest changes during the summer season in the western Mediterranean (∼11◦ C below modern), in the Adriatic (6◦ C below modern), and Aegean (7◦ C below modern), while winter temperature anomalies were less pronounced, with maximal decreases at 6–5◦ C below modern values recorded at Alboran, Gulf of Lions, and the Aegean (Hayes et al. 2005). Studies of the net freshwater budget suggest reduced LGM evaporation from the colder Mediterranean Sea (Chapter 2). With regard to climate modelling, several palaeoclimatic simulations are available for the LGM on a global scale, but perhaps of particular interest here are recent results using a nested model for the European sector (Barron and Pollard 2002; Pollard and Barron 2003). The model employs a global General Circulation Model (GENESIS2) to provide the lateral boundary conditions for a more detailed regional European climate model with a 60 × 60 km output grid. This fine grid allows the model to take into account topographic variability, leading to more accurate precipitation reconstructions than hitherto achieved. The uncertainty of the reconstructions is less than 3◦ C for monthly mean temperatures and ∼20 per cent for mean monthly precipitation. The LGM experiment shows that southern Europe was 7–10◦ C cooler than the modern simulation. Annual precipitation was reduced compared to present, especially during the winter (up to 4–8 mm per day in places in south-eastern Europe), while summer precipitation was somewhat reduced in the western Mediterranean and changed little in the eastern Mediterranean (Barron and Pollard 2002; Barron et al. 2004). The temperature reconstructions are in agreement with independent evidence from rock glacier distributions in the Pindus Mountains suggesting 8–9◦ C temperature depression during the LGM (Hughes et al. 2003; see also Chapter 12). Thus, the more recent evidence emerging from the Mediterranean appears to be in agreement in suggesting arid and cold conditions for the LGM. The main outstanding source of apparent discrepancy is the evidence of the LGM megalakes from the Dead Sea Transform area. Here, however, recent work supported by 14 C and U-series dating of sediments of Lake Lisan (the precursor of the Dead Sea) has produced a much more detailed view of the timing of lake-level changes during the last glacial (e.g. Bartov et al. 2002). During most of MIS 4 and MIS 3, the lake was ∼125 m above the present Dead Sea level. At ∼27 ka the lake rose sharply, reaching maximum values of over 200 m above present levels at about 25 ka and started dropping after 24 ka.

104

Chronis Tzedakis

Between 22 and 17 ka lake levels were about 180 m above the present Dead Sea level and then dropped considerably (Bartov et al. 2002, 2004). 6 A very similar history of lake-level changes has been reconstructed for Lake Kinneret (Sea of Galilee) to the north and indeed during the time of maximum expansion (26–24 ka), it appears that the two palaeolakes merged into a unified water body (Hazan et al. 2005). Thus, while the evidence for extensive lakes during the last glacial in the Near East appears incontrovertible, the precisely reconstructed temporal framework shows that maximum lake levels were reached during the interval 27–24 ka. During the LGM, lake levels were still significantly higher than present, but the important point is that the overall trend after 24 ka was towards reduced lake levels (allowing for short-term millennial-scale fluctuations— see mesoscale section). What emerges from this discussion is that the controversy surrounding the LGM climate of the eastern Mediterranean may have been somewhat exaggerated and that higher-resolution analyses and improved chronologies have led to a revised framework of geographically coherent changes. This effectively removes the need to invoke climatically complex scenarios and special pleading to reconcile apparently conflicting evidence. At the same time, however, it shifts the debate to an earlier interval and poses new questions over the origin of the high effective moisture in Israel during the period 27–24 ka. Examination of Figure 4.6 shows that this interval was characterized by prolonged interstadial conditions, with increased SSTs in the Mediterranean and expanded temperate tree populations in Greece. In addition, five new AMS radiocarbon dates from the Kastritsa Cave at Ioannina (Galanidou et al. 2000; Galanidou and Tzedakis 2002) place the beach deposits between 24.5 and 28.5 ka (using the Fairbanks et al. 2005 calibration). However, before complex climatic scenarios are once more invoked, there is an immediate need for new high-resolution records supported by a detailed chronological framework from across the Mediterranean to establish the palaeoenvironmental setting for this interval. 4. Tempo of Mediterranean climate variability. A certain degree of confusion exists over the fundamental pacing of Mediterranean climate change at different times in its Cenozoic history. For example, a recurring theme that appears in the literature is that prior to 2.8 Ma, precession was the dominant climate rhythm, but this does not 6

Bartov et al. (2002, 2004) consider the interval 24–7 ka as part of MIS 2, but in fact it belongs to the late MIS 3 of the SPECMAP stratigraphy (e.g. Martinson et al. 1987).

apply to every aspect of climate variability. With regard to global ice volume changes, as discussed above, obliquity was the dominant beat of the glaciated part of the Cenozoic and ‰18 Obenthic records vary primarily at the 41 kyr period, including the interval before and after the onset of major NH glaciation. On the other hand, subtropical North African climate, as recorded by the pattern of aeolian dust deposition in Atlantic and Arabian Sea marine cores, varied at the 23–19 kyr period before 2.8 Ma, reflecting the influence of precessional forcing of monsoonal climate (deMenocal 2004). After 2.8 Ma, the NH glacial tempo becomes apparent in the dust record with variability shifting to the 41 kyr period, increasing in amplitude after 1.7 (±0.1) Ma and then shifting to the 100 kyr period after 1 (±0.2) Ma. This evolution in subtropical North African climate variability is coeval with changes in the onset and amplification of NH ice sheets and points to a coupling between high- and low-latitude climates (deMenocal 2004). In the Mediterranean, the influence of obliquity is seen in ‰18 Oplanktonic and SST records over the last 5.3 Ma, becoming amplified from 2.8 Ma onwards, and reflects a high-latitude control (Lourens and Hilgen 1997). In addition, sapropel and dust deposition patterns over the past 12 million years (Lourens et al. 1996, 2001) reveal that obliquity also has a direct effect on circum-Mediterranean climate through dry–wet oscillations which are independent of glacial– interglacial variability (Tuenter et al. 2003), though the extent to which this represent a high-latitude signal is unclear. On the other hand, the same records clearly indicate the pervasive influence of precession throughout the last 5.3 Ma, being dominant prior to 2.8 Ma and continuing to persist after that into the interval of obliquity dominance. As explained earlier, the ‰18 Oplanktonic record is closely related to sapropel deposition and by extension to African monsoon variability and runoff from North Africa into the Mediterranean, and as such it is a low-latitude signal. However, the precessional variability in SST reflects the direct influence of insolation (Lourens et al. 1992) at the Mediterranean latitudes, rather than indirect African or high-latitude effects. Thus, as in subtropical Africa (as elegantly articulated by deMenocal 2004), precession appears to be a fundamental pacing of Mediterranean climate, but highlatitude glacial effects become superimposed on this signal after 2.8 Ma.

Plio-Pleistocene Glacial–Interglacial Cycles From the onset of major NH glaciation, ∼2.8 Ma to about 0.92 Ma, maximum ice volumes ranged between

Cenozoic Climate and Vegetation Change

105

(a)

_

(b) _

_

(c) _

(e) (f)

Fig. 4.6. (a) Variations in ‰18 O composition of benthic foraminifera in V19-30 in the East Pacific (Shackleton et al. 1983). (b) Variations in ‰18 O composition of ice in the GISP2 (Greenland Ice Sheet Project 2) record (Stuiver and Grootes 2000). (c) Variations in alkenone-derived sea surface temperatures in marine core MD95-2043 from the Alboran Sea, western Mediterranean (Cacho et al. 1999). (d) Interval of maximum lake levels of Lake Lisan, Dead Sea Transform area (Bartov et al. 2002, 2004). (e) Interval of Kastritsa beach deposits, Ioannina basin, north-west Greece (Galanidou et al. 2000; Galanidou and Tzedakis 2002). (f) Temperate tree pollen percentage curves from Ioannina I-284 (solid line), Kopais K93, central Greece (dotted line) (see Figure 4.8) (Tzedakis et al. 2002b).

Chronis Tzedakis

_

106

Fig. 4.7. June insolation for 65◦ N (Berger 1978) and variations in ‰18 O composition of benthic foraminifera over the last 3 Myr in the Shackleton 06 (S06) composite record from sites in the equatorial East Pacific: 0–341 ka from V19-30 (Shackleton et al. 1983), 341–430 ka from V19-28 (Ninkovich and Shackleton 1975), 431–98 ka from V19-25 (Shackleton unpubl.), 501–1,809 ka from ODP 677 (Shackleton et al. 1990), 1,812–3,000 ka from ODP 846 (Shackleton 1995a, b) (Data available from ). The interval of the Middle Pleistocene Transition (MPT) and the Mid-Brunhes Event (MBE) are shown. Glacial marine isotope stages discussed in the text are indicated.

one-half and two-thirds of the LGM value (Shackleton 1995). The first extensive glaciation is recorded during MIS 22, centred around 0.9 Ma (Shackleton et al. 1990). Between 0.92 and 0.64 Ma (the ‘Mid-Pleistocene Transition’ (MPT)), the climate system shifted to higher amplitude, lower-frequency variability characterized by the intensification of glaciation and a prolongation of the glacial–interglacial cycle (e.g. Mudelsee and Schulz 1997) (Figure 4.7). From 0.64 Ma onwards the 100 kyr glacial cycle became established, with most pronounced glaciations occurring at MIS 16, 12, 6, and 2 (Shackleton 1987; 1995; Shackleton et al. 1990). Another prominent shift is the ‘Mid-Brunhes Event’ (MBE) at the MIS 11/12 transition ∼430 ka, which marks the onset of maximum contrast between glacial and interglacial extremes (e.g. Berger and Wefer 2003). With respect to the terrestrial record, favourable geological conditions in southern Europe have in some cases

led to the relatively undisturbed accumulation of thick Quaternary sedimentary sequences. Such sequences provide an opportunity to develop complete records of terrestrial events over multiple glacial–interglacial cycles. The linking of the longest pollen sequences from southern Europe has led to the emergence of a coherent stratigraphical framework of changes in vegetation for the last 450 kyr and has allowed comparisons with the marine isotopic record (e.g. Tzedakis et al. 1997; 2001). This showed that the many stages and substages into which the marine isotopic sequence is divided are also appropriate for viewing the continental record, although the marine and terrestrial boundaries may not be precisely synchronous (e.g. Tzedakis et al. 2003c, Tzedakis et al. 2004b; Tzedakis 2005). These sequences and other more fragmentary Pliocene and Pleistocene records (see review of earlier records by Suc and Popescu 2005) essentially show that during the past 2.8 Myr, Mediterranean vegetation has oscillated between two extreme

Cenozoic Climate and Vegetation Change

situations: discontinuous herbaceous communities during glacials and interglacial forest/woodland. Between these two states, transitional phases occurred of varying duration and extent, depending on the direction and rate of change of the system and geographical location. Given that the availability of palaeoenvironmental information increases with decreasing age, an attempt will be made here to focus on the most recent glacial and interglacial maxima as case studies of vegetation responses. These are then compared with earlier periods and differences and similarities are discussed.

Glacials Biome distributions in southern Europe and Africa have been reconstructed for the LGM (Elenga et al. 2000). A ‘biomization’ procedure was used to convert pollen spectra to biomes and this was applied to eighteen records from southern Europe and the Near East (no LGM records were available in the databases from North Africa). The dominant vegetation was reconstructed as steppe (grassland or shrubland), with varying proportions of Artemisia, Gramineae, and chenopods. Scanning electron microscopy on chenopod pollen from Tenaghi Philippon, north-east Greece, showed the presence of Eurotia ceratoides and Kochia laniflora, which are found today in steppe and semi-desert environments in central Asia and indicate cold and arid conditions (Smit and Wijmstra 1970). A marine pollen record off north-west Africa shows semi-desert pollen types during the LGM (Agwu and Beug 1984) in agreement with the simulated extension of semi-desert in the southern Maghreb (Jolly et al. 1998). Refugia. Ever since the first full-glacial pollen diagrams from southern Europe (e.g. Wijmstra 1969; Florschütz et al. 1971) showed steppe-dominated landscapes, implying that forest biomes had not simply shifted en masse southwards as ice sheets expanded, the whereabouts of temperate elements of the flora has been a topic of continuous discussion. The prevailing hypothesis has been that remnant tree populations found refuge in the southern peninsulas of Europe where they survived in suitable microhabitats in midaltitude zones and in locally moist sites in lowland and coastal areas (e.g. Beug 1968, 1975; Frenzel 1968, 1979; Lang 1970; van der Hammen et al. 1971; Bennett et al. 1991; Birks and Line 1993). Given the small size of such populations and the relative lack of fullglacial pollen diagrams, direct palaeobotanical detection has been difficult. West (1980) suggested that information on the glacial locations of tree populations could

107

be obtained through examination of their postglacial migration routes. Huntley and Birks (1983) followed that suggestion by making the first European compilation of pollen records 0–13,000 14 C yr BP and, on the basis of the locations of first expansion of tree populations, proposed seventeen refugial areas in the closing parts of the last glacial stage. In the compilation of LGM records from Southern Europe and the Near East (Elenga et al. 2000), there is evidence of the local presence of trees, suggested by continuous or semicontinuous pollen curves of Pinus, Juniperus, Cedrus, Quercus in a number of sites. Perhaps the record that has been most closely linked to a classic refugial site in the Mediterranean is that from the Ioannina basin (Figure 4.9), situated in a topographically diverse landscape on the western flank of the Pindus Mountain range, Greece (Bottema 1974; Tzedakis 1993; Tzedakis et al. 2002b). A long pollen sequence spanning multiple glacial–interglacial cycles has provided evidence for the continued presence of temperate trees during the glacial periods, while the more thermophilous taxa show intermittent presence (Tzedakis 1993). A new highresolution record from the same basin shows that during the LGM, mean AP values were 35 per cent (min. 24%; max. 54%) and the record suggests the presence of deciduous Quercus, Abies, Pinus, Betula, Fagus, Ulmus, Carpinus betulus, Ostrya, and Alnus within the Ioannina catchment (Tzedakis et al. 2002b; Lawson et al. 2004). In view of this evidence, it may be worth considering what local intrinsic properties make the Ioannina area particularly well suited to the survival of refugial populations. During the LGM, tree growth in the Mediterranean would have been limited by: (1) increased aridity; (2) lower atmospheric CO2 content, which in turn leads to reduced water-use efficiency of plants, exacerbating water stress (Cowling and Sykes 1999); and (3) minimum winter temperatures (e.g. Larcher 1981). The impact of these climatic factors on vegetation communities was determined by the extent to which these changes crossed ecological thresholds controlled by local intrinsic properties (Tzedakis et al. 2002b). The persistence of tree populations at Ioannina during the last climate cycle is seen as a function of continued moisture availability from the nearby Ionian Sea, counteracting the effects of increased aridity and reduced CO2 concentrations. An additional factor is the degree of topographical variability (Figure 4.8), which determines the extent to which populations can shift altitudinally in response to climate change. Given that both temperature and CO2 concentration increase with decreasing altitude, vertical migration allows populations, at least partly, to evade

108

Chronis Tzedakis

Mount Tymphi (2497 m) Klithi rockshelter

Vikos Gorge

Mount Mitsikeli (1810 m)

Lake loannina (Pamvotis)

Kastritsa rockshelter

0

5 km

Fig. 4.8. SPOT imagery of the Ioannina basin and surrounding areas showing the extent of topographical variability in the region. The location of Mount Tymphi—which has one of the best glacial records in the Mediterranean (Chapter 12)—is also shown.

extirpation, provided that there is sufficient topographical variability to supply a range of microclimates suitable for survival. Thus the synergy of these local factors appears to have buffered the most extreme effects of Quaternary climate variability, allowing the persistence of tree populations. The Ioannina setting of a mid-altitude site within an area of varied topography and high moisture availability can thus be used as a type-site that can help us identify potential refugial areas around the Mediterranean. A study of topographical and precipitation maps (moving clockwise around the Mediterranean) would immediately point towards areas in the Sierra Nevada, Sierra de Segura, the Pyrenees, the southern part of the Maritime and Ligurian Alps, the western

flank of the Apennines, southern Apennines, the western flank of the Dinarides and Hellenides, Mount Olympus, the southern part of the Taurides, the Ansariye and Lebanon Mountains, the Maritime Atlas of Algeria, and the Atlases in Morocco. Some palaeoecological evidence is indeed available showing the presence of refugial populations in these areas (see Table 4.3 and Figure 4.10). Large Mediterranean islands of varied topography must have also harboured refugial populations, but here the palaeoecological record does not extend back to the LGM. The Cantabrian Mountains of northern Spain, the western (east Bulgaria) and southern (Pontus Mountains) Black Sea region, and the Caucasus were also important refugial areas, but these are beyond the Mediterranean realm.

Cenozoic Climate and Vegetation Change

109

TABLE 4.3. A sample of Mediterranean pollen records, spanning all or part of the LGM Site 1. San Rafael, Almería, Spain (36◦ 46 25 N, 2◦ 36 5 E, sea level) 2. Padul, Granada, Spain (37◦ N 3◦ 40 E, 785 m a.s.l.) 3. Siles, Sierra de Segura, Spain (38◦ 24 , 2◦ 30 ; 1,320 m a.s.l.) 4. Banyoles, Catalonia, NE Spain (42◦ 7 57 N, 2◦ 45 47 E, 173 m a.s.l.) 5. La Borde, eastern Pyrenees, France (42◦ 32 N, 2◦ 05 E, 1,660 m a.s.l.) 6. Bouchet (D), Massif Central, France (44◦ 55’N, 3◦ 47 E; 1,200 m a.s.l.) 7. Lagaccione, central Italy (42◦ 34 N, 11◦ 51 E, 355 m a.s.l.) 8. Valle di Castiglione, central Italy (41◦ 53 30 N, 12◦ 45 35 E; 44 m a.s.l.), 9. Lago Grande di Monticchio, southern Italy (40◦ 56 40 N, 15◦ 36 30 E, 656 m a.s.l.) 10. Ljubljana Moor, Slovenia (45◦ 59 N, 14◦ 25 E; 300 m a.s.l.) 11. Tenaghi Philippon, north-east Greece (41◦ 10 N, 24◦ 20 E; 40 m a.s.l.) 12. Ioannina 284, Epirus, Greece (39◦ 40 N, 20◦ 51 E; 470 m a.s.l.) 13. Xinias, central Greece (39◦ 03 N, 22◦ 16 E; 480 m a.s.l.) 14. Kopais, central Greece (38◦ 26 N, 23◦ 03 E; 95 m a.s.l.) 15. Ghab, north-west Syria (35◦ 41 N, 36◦ 18 E, 240 m a.s.l.) 16. Dar Fatma 2, Kroumirie Mts, Tunisia (36◦ 50 N8◦ 45 E, 780 m a.s.l)

17. Tigalmamine, Middle Atlas, Morocco (32◦ 54 N, 5◦ 21 W, 1,628 m a.s.l.)

LGM arboreal pollen spectra Olea, Pistacia, evergreen and deciduous Quercus, Vitis, Corylus Pinus, deciduous Quercus, Juniperus, traces of evergreen Quercus Pinus, deciduous and evergreen Quercus, Corylus, Betula, Fraxinus Pinus, traces of Juniperus, Betula, Quercus (deciduous and evergreen), Olea Pinus, Abies, traces of Betula, deciduous Quercus, Corylus, Fagus, Carpinus Pinus, traces of Betula and Juniperus Pinus, Juniperus, traces of deciduous Quercus, Corylus, Betula, Salix Pinus, yJuniperus, some deciduous Quercus and Salix, traces of Corylus, Fagus, Ulmus, Betula Pinus, Juniperus, Betula, some deciduous Quercus, traces of Ulmus, Ostrya Pinus, Picea, Betula

References Pantaléon-Cano et al. 2003 Pons and Reille 1988 Carrión, 2002 Pérez-Obiol and Julià 1994 Reille and Lowe 1993 Reille and de Beaulieu 1990 Magri 1999 Follieri et al. 1988 Allen et al. 2000 Sercelj 1966

Pinus, Juniperus, Betula, Alnus

Wijmstra 1969

deciduous Quercus, Abies, Pinus, Betula, Fagus, Ulmus, Carpinus betulus, Ostrya, and Alnus Pinus, deciduous Quercus, Betula, Juniperus and Abies, traces of evergreen Quercus, Pistacia, Ulmus, Ostrya and Corylus deciduous Quercus, Pinus, Juniperus, Abies, traces of evergreen Quercus and Pistacia deciduous Quercus, Pinus, Juniperus, Cedrus, Salix, traces of Ostrya, Fraxinus, Olea, Pistacia, Phillyrea Quercus canariensis (deciduous) in high abundance, Alnus, Pinus, Abies, Cedrus, Quercus suber, Juniperus, traces of evergreen Quercus, Ulmus, Olea, Salix deciduous Quercus in low abundance

Tzedakis et al. 2002b Lawson et al. 2004 Bottema 1979

The absence of mediterranean sclerophylls during the LGM from sites such as Ioannina suggests that in addition to the moist mid-altitude refugial sites, another type of refugium was also present located in low-altitude (valley bottoms and coastal plains) and lower latitude areas which were warmer but also drier. However, direct palaeobotanical evidence for such refugia is not in great abundance (but see the site of San Rafael in southernmost Spain: Pantaléon-Cano et al. 2003). Eustatic sealevel rise would have drowned the coastal plains that were exposed during the LGM lowstand in parts of the Mediterranean and with them the evidence of refugial stands (e.g. Triat-Laval 1978). In addition, the distribution of trees at low-altitude sites may have resembled a savanna, with scattered low-density populations. This would be represented in pollen diagrams as low

Tzedakis 1999; Okuda et al. 2001 Niklewski and van Zeist 1970 Ben Tiba and Reille 1982

Lamb et al. 1989

(background) percentages, which are difficult to interpret and often dismissed as ‘long-distance transport’. Moreover, some of the mediterranean taxa are palynologically underrepresented, which may account for their minimal and intermittent palynological values and the domination of the glacial arboreal pollen curves by taxa such as Pinus and Juniperus (see low-altitude sites in Table 4.3). The advent of modern molecular techniques for studying the genetic diversity of modern populations has recently provided significant insights into species lineages, migration routes, and refugial locations (see reviews by Hewitt 1996, 1999, 2000). A number of studies on plant and animal species sampled across their range are now beginning to emerge from Europe (e.g. Taberlet et al. 1998; Hewitt 1999; Petit et al. 2002) and

110 Chronis Tzedakis

10 6 5 4

7 8 11 9 12

3 13

2

14

1

15

16 17

Tree types Corylus

Pinus

Alnus

Fagus

Betula

Juniperus

Quercus (deciduous)

Abies

Cedrus

Mediterranean

Salix

Carpinus

Picea

Ulmus

0

500 km

Fig. 4.9. Location of some Mediterranean pollen records from wetland sites spanning all or part of the LGM, with inferred refugial tree populations (only main tree taxa are shown; for more information refer to Table 4.3). Site numbers correspond to those in Table 4.3.

Cenozoic Climate and Vegetation Change

111

0

Age (Ma)

1

2

3

r ba am id qu Li va a o lk ary Ze oc n er a r o Pt roti n d r e Pa l l o d he a P ay ia h at m C om s su uc E a cis y o ar n C e ys rth pit Pa do a ci S ga e a e u c a Ts o t a r d i ap a n S elh dro ng n E de r i o um L i odi x a i Ta u o ca eq S sa na i ys p N ser a ro li P no ag e M yal ur ia E id in a ct ti A ar rix w a te ol S ud a se m P i o s k ia el s M win ar K sia um e n al m H pre s pi yro sp io E

D

Fig. 4.10. ‘Serial extinction’ of a number of genera in Europe. Zelkova and Liquidambar are shown as persisting today, although they have disappeared from mainland Europe and are found only on Mediterranean islands. Modified from van der Hammen et al. (1971).

these also indicate that the southern European peninsulas were the source of most of the postglacial colonization of Europe. (Hewitt 1996, 1999, 2000) envisages a ‘leading edge’ expansion from populations at the northern limits of the refugial areas, with longdistance dispersers establishing colonies far ahead of the main population and rapidly expanding to fill the area before others arrived. One implication of the leading edge expansion is that refugial populations further to the south may have been blocked and therefore may have had a much reduced contribution to the postglacial recolonization of northern Europe (Hewitt 2000). This predicts that populations from the southern part of the peninsulas would have distinct genetic signatures not found in northern Europe. This can be seen in the genomic structure of Alnus glutinosa (King and Ferris 1998) with unique haplotypes in Greece and Bulgaria, while the rest of northern Europe shares haplotypes originally from the northern Balkans. However, establishing the northernmost extent of glacial refugia in Europe is far from straightforward because a species’ distributional pattern is likely to comprise larger southern populations with progressively smaller northern peripheral ones (Hewitt 1996). Moreover, postglacial expansion may include advances and retreats due to climatic oscillations (like the Younger Dryas) and thus northern remnants of an earlier advance may form the ‘secondary’ refugia for a subse-

quent wave of expansion (Brewer et al. 2002). Separating the effect of these events and resolving the location of the ‘northern’ or ‘leading edge’ during glacial maxima, therefore, requires spatially detailed genetic and fossil evidence, which is still largely unavailable. On the basis of the fossil and genetic evidence outlined above, the following generalizations on the distribution of refugial tree populations in southern Europe can be made. Populations of Pinus, Picea, Larix, Betula, Salix, and Juniperus were present in the northernmost part of the southern European peninsulas and indeed further north during the LGM. With the possible exception of Alnus and Corylus, the LGM northern edge of most temperate trees would have been south of 46◦ N, probably even further to the south within hinterlands away from coastal areas. Populations of mediterranean species (Olea, Pistacia, Phillyrea, evergreen Quercus), whose distributions are limited by minimum temperatures (Larcher 1981), would have been further to the south still. This view has been recently challenged by the notion of ‘northern cryptic refugia’ of temperate trees with seriously disjunct distributions north of the Iberian, Italian, and Balkan peninsulas (Stewart and Lister 2001; Stewart 2003) and by the suggestion that during the full glacial, central and eastern Europe were partly occupied by a taiga/montane woodland, with pockets of temperate trees (Willis et al. 2000; Willis and van Andel 2004). The main evidence for these claims is macroscopic

112

Chronis Tzedakis

charcoal, often from archaeological settings. However, only a subset of the charcoal pieces has been directly dated, usually charcoal pieces of Pinus or Picea. The rest have associated dates from the same units, which in view of the problems with remobilized charcoal raises some concern over their chronological integrity. The main problem, however, is the definition of the temporal interval of interest as it pertains to the survival of refugial populations. The charcoal data span a very large time window (42–19 ka), which conflates intervals of quite distinct environmental conditions. Instead, a strict consideration of the refugial distribution of trees should be confined to the LGM (i.e. 21 ± 2 ka), reflecting the most extreme interval in Europe in terms of ice expansion and CO2 depression, which is when population size and genetic diversity reach their minima and species survival hangs in the balance. When this is done a very different picture emerges in that the only macrofossils present are from Pinus, Picea, Larix, Salix, and Betula (see Willis and van Andel 2004: Figure 4.2). As regards the phylogeographical information, no genetic evidence exists for disjunct refugia of temperate trees far north of the southern European peninsulas, apart from Corylus avellana and Alnus whose genomic structure suggests that postglacial recolonization of Europe originated from outside the classic refugial peninsulas (Palmé and Vendramin 2002; Petit et al. 2003). Only genetic evidence from boreal tree species (Picea abies, Betula, Populus, Salix) suggests more northerly refugial distributions and postglacial recolonization from intermediate latitudes (Lagercrantz and Ryman 1990; Palmé et al. 2003a, b; Petit et al. 2003; Lascoux et al. 2004). By contrast, molecular markers of Fagus (Demesure et al. 1996), Quercus (Petit et al. 2002), and Carpinus (Grivet and Petit 2003) show that the highest levels of variation are found in southern refugial peninsulas. This is in agreement with the palynological evidence presented earlier, which shows a clear latitudinal distinction between boreal and temperate tree species in terms of their glacial distributions. Serial extinctions. The term ‘species extinction’ covers a broad range of events. At one end of the spectrum are mass-extinctions, occurring on a global scale over a relatively short period of time. At the other end of the spectrum are localized disappearances of individual species. Somewhere in between are ‘serial extinctions’ affecting several species over a broad geographical region and occurring over a relatively long time span. Such a gradual disappearance of different types of trees occurred in Europe over the last ∼2.8 Ma and led to significant floristic changes (Figure 4.10). This has resulted in a presentday European tree flora that is significantly reduced in

taxonomic diversity compared to the Early Pliocene, and also compared to the diversity of North America and Asia, where all the extirpated European tree genera are still present. The continuous disappearance from Europe of various tree species has been explained as a function of the east–west orientation of mountain ranges which presented a significant obstacle to southward migration (Reid 1935): the idea was that some species retreating from northern Europe at the onset of glaciation were unable to cross the mountain barriers and consequently became extinct. Such cases of failed retreat were repeated during each glaciation and this could account for the ever-growing list of casualties in Europe. By contrast, in eastern North America, where so many more tree species have survived the Quaternary ice ages, mountain ranges are oriented north–south and therefore did not present a significant barrier to migration. However, after examining the available late interglacial and early glacial pollen diagrams from central Europe, Bennett et al. (1991) found no evidence for a southward migration of temperate tree populations. This, in conjunction with the observation that at present most of the north European temperate tree species are also found in southern European mountains, led Bennett et al. (1991) to suggest that during the population contraction phase at the onset of a stadial or glacial there is no reverse movement towards southern refugial areas, but rather northern populations of temperate trees degrade in situ. On the other hand, the reappearance of Pinus and Betula during the closing phases of interglacial vegetation successions, means that the possibility of the retreat of some boreal trees towards central Europe could not be excluded. In southern Europe, local tree populations would also contract, but a subset of them would be able to remain in suitable areas for survival. From these glacial stations temperate trees were able to expand during intervals of favourable climate conditions, but a part of the population would always remain in the south, providing the long-term continuity. Thus, the maintenance of temperate tree species in Europe becomes inexorably linked to the long-term persistence of part of the populations in the south. Northern populations of most temperate species are irrelevant to their long-term survival because they disappear at the end of interglacials; it is the failure of a species’ southern populations to survive during either an interglacial or a glacial interval that ultimately leads to its extinction from Europe. A corollary of this is that the east–west orientation of mountains in Europe is unlikely to have led to the disappearance of many Tertiary species by acting as a barrier to spread. Thus, the enhanced rate of extinction in Europe relative

Cenozoic Climate and Vegetation Change

to eastern North America and Asia is seen as a function of the much-reduced area available for survival of refugial populations south of the Alps (Bennett et al. 1991; Huntley 1993). Instead of being largely a matter of chance, the PlioPleistocene extinctions of tree species in Europe appear to have been deterministic. West (1980) drew attention to the fact that most of the Tertiary genera which became extirpated in Europe are monotypic and have restricted present ranges in other continents, while those with large variability have been consistently successful in surviving the environmental oscillations. In an insightful analysis of the climatic requirements of European Tertiary and Quaternary cool-temperate tree flora, Svenning (2003) found that genera that are still widespread in Europe are the most tolerant of cold growing season and winter temperatures, while genera now restricted to the Mediterranean area are cold-sensitive but relatively drought-tolerant. The least tolerant genera are now extinct from Europe, but survive in regions of North America and Asia. The analysis also points towards the existence of two types of tree refugia in southern Europe, in agreement with inferences from the palaeobotanical information discussed earlier: (1) cool, moist mid-altitude sites; and (2) a less common type, of warmer and more drought-prone low-altitude sites. Accordingly, the widespread genera were sufficiently cold-tolerant to occupy the mid-altitude sites, while the genera presently restricted to the Mediterranean area (relictual taxa, sensu Svenning 2003) survived at lowaltitude sites. The relative scarcity of low-altitude refugia coupled with their reduced likelihood to buffer environmental oscillations, may have led to increased vulnerability of the relictual taxa, which could account for their presently restricted distribution. Although not directly discussed by Svenning (2003), his analysis also provides an explanation for the diachronous disappearance of taxa from north to south during the course of the Plio-Pleistocene. Palaeobotanical evidence from across Europe suggests that while a number of thermophilous elements were extirpated from north-west Europe after the first glacial, they persisted for much longer in southern Europe. For example, Taxodiaceae disappeared after the Praetiglian Stage (c.2.8 Ma) in the north (e.g. van der Hammen et al. 1971; de Jong 1988), but survived in the south until ∼1 Ma (Suc and Popescu 2005) and in Crotone (Calabria, Italy) were recorded up to MIS 11 (Capraro 2002). Pterocarya is considered to have disappeared in Europe after MIS 11, but in the Massif Central (M. Reille, pers. comm.) and at Tenaghi Philippon, northeast Greece (Smit 1976) its last appearance was in MIS 9, while at Valle di Castiglione, central Italy, it persisted

113

until MIS 7 (Follieri et al. 1988). Finally Zelkova has disappeared from continental Europe, but has survived in small populations in Sicily and Crete. Using Svenning’s (2003) framework and terminology, it appears that during the course of the Plio-Pleistocene, formerly widespread taxa in the Tertiary became first relictual and restricted to southern Europe, before becoming extinct. The question then is why in their relictual phase these taxa did not spread north during interglacials? Perhaps northern European interglacial climates were already too cold for them to survive. Alternatively, the restricted ranges of the relictual taxa may have led to a reduction of their genetic variability or adaptations that prevented them from migrating. In addition, their northward expansion might have been blocked by the widespread taxa, whose more northerly refugial distribution would have formed the ‘leading edge’ for the interglacial colonization of Europe. Earlier glacials. Tzedakis et al. (2003c) used the absolute minima of AP percentages at Tenaghi Philippon, north-east Greece, as a first-order index of the relative extremity of individual glacial stages and showed that the most severe tree population contractions of the last 450 kyr (in descending order of severity) occurred during the glacial maxima of MIS 12, 6, 2, 10, and 8. These results were in agreement with the relative extent of ice sheets as estimated from marine benthic isotope records (e.g. Shackleton 1987; Waelbroeck et al. 2002) and reconstructed from field evidence from northern Europe (Ehlers 1996), and indicated a close correspondence between size of tree populations and ice volume (Tzedakis et al. 2003c). In addition, recent evidence based on U-series assays for glaciation in the Pindus Mountains, north-west Greece, indicates that the most extensive glaciations in descending size were MIS 12, 6, and 2 (J. C. Woodward et al. 2004; Hughes et al. 2006; Chapter 12). These comparisons can now be extended further back in time by examining the lower part of the Tenaghi Philippon record and using a revised timescale based on pollen-orbital calibration procedure (Tzedakis et al. 2006). The new age model suggests that the base of the sequence had been significantly underestimated and extends it from 1 Ma back to 1.35 Ma. This provides an excellent opportunity to assess the character of glacial vegetation over a long interval at the same location and, more importantly, extending the comparisons well into the 41-kyr world of obliquity-dominated glacial cycles (Figure 4.11). Over the interval 0.45–0.9 Ma, the Tenaghi Philippon record again shows good correspondence with ice volume reconstructions, with the most extreme and sustained AP minima occurring during MIS 16 and also MIS 22, closely reflecting variations

114

Chronis Tzedakis 100-kyr state

3

41-kyr state

MPT

d18 Obenthic (‰)

S06 3.5 4 4.5 5 2

6

8

10

12

14

16

18 20

22 24 26 28 30

34

36 38 40 42

AP (%)

100

50

TP

0 0

200

400

600

800

1000

1200

1400

Age (ka) Fig. 4.11. Variations in ‰18 O composition of benthic foraminifera in the S06 composite record (see the caption to Figure 4.7) and arboreal pollen percentages (AP) at Tenaghi Philippon, north-eastern Greece (Wijmstra 1969; Wijmstra and Smit 1976; van der Wiel and Wijmstra 1987a, b) over the last 1,400 kyr. The new Tenaghi Philippon timescale is based on Tzedakis et al. (2006). Aquatics, Gramineae, and locally overrepresented taxa are excluded from the pollen sum. Glacial marine isotope stages are indicated.

in the amplitude of the benthic isotope curve (Shackleton et al. 1990). MIS 16 appears to be the most extensive glacial period of the last 1.35 Myr, not in terms of the absolute AP minimum (this is recorded in MIS 12), but rather of the prolonged suppression of tree populations. Prior to 0.9 Ma, maximum ice volumes ranged between one-half and two-thirds of the LGM value (Shackleton 1995). By analogy, therefore, one would expect that the tree population contractions would be less extensive. Instead, the Tenaghi Philippon record shows the occurrence of AP minima that are comparable to the Middle and Late Pleistocene: for example, non-arboreal pollen (NAP) values during MIS 38, 40, and MIS 42 exceed 90 per cent, with Artemisia and chenopods at over 80 per cent, indicating extreme aridity. The difference is that these Early Pleistocene AP minima are not sustained over long periods (<10 kyr), compared to the more temporally extensive AP minima of MIS 22, 16, 12, 6, and 2. One can speculate whether the shorter glacials led to a different geographical distribution and size of refugial populations of temperate trees during the Early Pleis-

tocene. Indeed, it has been argued that the less extensive glaciations of the Early Pleistocene glacial would have allowed temperate tree populations to survive in central Europe (e.g. de Jong 1988). However, the pattern of AP changes in the Early Pleistocene at Tenaghi Philippon suggests that even short glacial spells were severe enough to lead to major population contractions in southern Europe and by extension to central and northern Europe This is also seen in Late Pliocene records from France where large AP decreases are recorded during cold stages (e.g. the Sénèze pollen record from the Massif Central, showing extensive arboreal contractions during MIS 82: Elhaï 1969; Roger et al. 2000). It is thus difficult to see how forest collapses would be less severe in central and northern Europe than further south.

Interglacials Given the degree of spatial heterogeneity around the Mediterranean, the influence of local factors in producing a mosaic of environmental responses is an important

Cenozoic Climate and Vegetation Change

component of palaeoenvironmental reconstructions in the region (e.g. Magri et al. 2004). An attempt is made, therefore, to extract patterns of vegetation development from the most recent interglacials, providing information on the geographical complexity around the Mediterranean. Here, interglacial vegetation development is mainly reconstructed with reference to the few records from the Last Interglacial, while Holocene records are used to add regional detail, especially in cases where no Last Interglacial records are available. This initially may seem surprising since the Holocene record is significantly more voluminous and has the added advantage of an independent time control, mainly in the form of 14 C dating. However, this advantage may in fact be a handicap in disguise. This is because a great deal of 14 C dating has been undertaken on bulk samples from sediments, which, given the wide occurrence of carbonate terrains in the Mediterranean region, may be particularly prone to ‘hard-water’ effects. Thus, while there has been a tendency to emphasize apparently incongruent patterns, these are often an artefact of poor dating control and/or insufficient resolution (see e.g. discussions by RossignolStrick 1995; Meadows 2005). Indeed, the emergence of new high-resolution data sets with improved chronologies has revealed concordant regional relationships. Moreover, Holocene pollen records have the added complication of containing an anthropogenic signal, which in some cases extends over a good part of the interglacial and masks many natural changes (e.g. Yasuda et al. 2000). The concept of a glacial–interglacial cycle of vegetation and soil changes was first proposed for northwest Europe by Iversen (1958) and developed by Andersen (1966), Turner and West (1968), and Birks (1986). It comprises a glacial (cryocratic; open vegetation, immature soils), a pre-temperate (protocratic; pioneer vegetation, unleached soils), an early-temperate (mesocratic; development of closed forest, fertile soils), a late-temperate (oligocratic; decreasing forest cover, deteriorating soils) and a post-temperate (telocratic; increasingly open vegetation, infertile soils) phase. This scheme was modified for the Mediterranean by van der Hammen et al. (1971), such that a glacial steppe was succeeded by sclerophyllous woodland, then a deciduous mixed oak forest followed by a retrogressive phase. The term ‘interglacial vegetation succession’ refers to the recurring pattern of vegetational development during the course of a temperate stage, which is complex and progressive and reflects the reorganization of vegetation communities under the influence of climate forcing, internal biotic factors, soil changes, and historical aspects, including conditions during the preceding glacial stage (e.g. Birks

115

1986; Tzedakis and Bennett 1995). In general, southern European sites with sufficient moisture availability, show a pre-temperate phase of open woodland (Juniperus, Pinus, Betula, Quercus), then a temperate phase characterized by early expansion of deciduous Quercus, mediterranean sclerophylls, Ulmus, Corylus, and Tilia, followed by Carpinus, Ostrya, and then Abies and Fagus (and sometimes Picea); a post-temperate phase of open woodland ensues before the onset of glacial steppe. Closer examination of the available data reveals specific regional patterns in the character of the interglacial succession, reflecting climate and local intrinsic properties. Thus, during the Last Interglacial, Greek (Tzedakis et al. 2003b) and Italian sites (Follieri et al. 1988; Brauer et al. 2007) show deciduous Quercus and Ulmus expanding first, followed by an increase in mediterranean sclerophylls; about 3 kyr into the interglacial Carpinus and Ostrya populations expand; Abies then increases and remains dominant until the end of the interglacial at Lago Grande di Monticchio (Allen 2005), while at Ioannina the final phase of the interglacial shows a reexpansion of deciduous Quercus (Figure 4.5). In the more continental site of Tenaghi Philippon, the Last Interglacial succession is largely dominated by Quercus, with Pinus expanding towards the end (Wijmstra 1969; Wijmstra and Smit 1976). Further to the west, in the Massif Central, the sites of Bouchet (Reille et al. 1998) and Ribains (de Beaulieu and Reille 1992) show both central European and Mediterranean affinities, thus conveniently linking the stratigraphies north and south of the Alps. After peaks in Pinus and Betula, the full interglacial starts with deciduous Quercus, but mediterranean sclerophylls are largely missing (although long-distance traces are still visible, probably reflecting expansion at low altitudes). Corylus expands shortly after Quercus and forms a major part of the early succession, while an expansion of Taxus is also notable. This is followed by a major expansion of Carpinus and then Abies. Picea is an important part of the later stages of the succession, reflecting the sites’ high altitude and geographical position. The interglacial ends with Pinus and Betula. Comparison between Last Interglacial and Holocene diagrams from the same regions shows the following differences: mediterranean sclerophylls had a larger expansion during the early part of the Last Interglacial than in the early Holocene, a reflection of the higher summer insolation maximum centred at 127 ka. Corylus in the west and Quercus in the east assumed larger roles in the early Holocene. Carpinus was a characteristic taxon of the Last Interglacial succession at most sites, while during the Holocene Ostrya assumed a more dominant role in the east and Tilia and Quercus in the west; Fagus was

116

Chronis Tzedakis

largely absent during the Last Interglacial from southern (and indeed northern) Europe, with the exception of remnant populations in Italy (see discussion in Tzedakis 1994). In contrast to the diverse interglacial successions observed at sites from north-east Spain, southern France, northern and central Italy, and northern Greece, sites from southernmost Europe and from generally drier locations record Holocene vegetational developments that are mostly dominated by deciduous and evergreen Quercus, other mediterranean sclerophylls, and by Pinus and Juniperus (e.g. Pons and Reille 1988; PantaléonCano et al. 2003; Sadori and Narcisi 2001; Bottema and Sarpaki 2003). In fact a clear succession pattern is often difficult to discern, but this is also a function of anthropogenic effects altering the natural vegetation succession (e.g. large increases in Olea during the second half of the Holocene, which are certainly not a feature of the usual succession). Moving to the Near East, Holocene diagrams from south-central and southwestern Anatolia (see review by van Zeist and Bottema 1991), Syria (Niklewski and van Zeist 1970), and Israel (Baruch et al. 1999) are dominated by deciduous and evergreen Quercus, Juniperus, Cedrus, Pistacia, Olea, and Ostrya, the relative abundances of these taxa changing with location. A striking feature of these records is that the expansion of woodland does not occur at the Pleistocene/Holocene transition of 11.5 ka, but is often delayed until about 8 ka (e.g. Wick et al. 2003). This may be a function of the increased and extended aridity during the boreal summer insolation maximum, the degree of the delay depending on local conditions keeping moisture availability below the tolerance threshold for tree growth. Another feature that emerges from many sites across southernmost Europe and the Near East, is that AP maxima do not exceed 50–60 per cent, suggesting that closed forest conditions were never established in these areas during the Holocene. It is possible that this may be an artefact of anthropogenic disturbance especially in the eastern Mediterranean (e.g. Yasuda et al. 2000), but similar features are seen in the western basin from the onset of the Holocene (e.g. Pons and Reille 1988; Pantaléon-Cano et al. 2003). Records from earlier interglacials in the same areas are needed to settle this matter conclusively. On the North African coast, pollen records from marine cores in the Nile cone area show large expansions of Quercus, Olea, Pistacia, and other mediterranean taxa coeval with the deposition of sapropels S5 and S1 (Cheddadi and Rossignol-Strick 1995). Further to the west, in the Kroumirie Mountains of Tunisia, the early Holocene is characterized by diverse communities

including deciduous and evergreen Quercus, Pinus, Cedrus, but after 6–5 ka a distinct increase in herbs is recorded, with the reduced arboreal component represented mainly by Quercus suber and Erica arborea (Ben Tiba and Reille 1982). In the Akfadou coastal mountains, north-east Algeria, the pollen record from La Châtaigneraie peat bog (1,225 m a.s.l.) shows that Cedrus, which completely dominated the vegetation (values reaching 90%) during the Late Glacial, was gradually replaced in the early Holocene by deciduous Quercus, while the appearance of Quercus suber suggests its expansion at lower altitudes (Salamani 1993). Interestingly, Holocene AP values did not exceed 70 per cent, and in the earliest Holocene were usually ∼50 per cent, but this appears to be a function of the high percentages of Gramineae, which may represent a local semi-aquatic signal. By comparison, the record from Tigalmamine, in the Middle Atlas of Morocco, shows the dominance of evergreen Quercus throughout the Holocene, with Cedrus arriving about 5 ka (Lamb et al. 1989). Equatorwards from the North African coast, early Holocene lake sediments from the Sahara have provided evidence of increased moisture availability over the interval 10– 6 ka and palynological results show that savanna and desert grassland extended over regions that today are occupied by hyperarid deserts (e.g. Ritchie and Haynes 1987). Indeed, palaeobotanical data suggest that the Sahara/Sahel boundary had migrated northwards at least as far as 23◦ N and that the Sahara was drastically reduced, under the influence of the strengthened African monsoon (Jolly et al. 1998; Prentice et al. 2000). Evidence from earlier interglacials is available mainly from southern Europe. This shows that over the last 430 kyr, the character of the vegetation succession was generally similar between interglacials, with a pool of taxa always expanding early, while others expanded in later phases (e.g. Tzedakis et al. 2001). Early Middle and Early Pleistocene interglacials again show a sequential expansion of taxa but with the increasing presence of the ‘Tertiary relicts’ as one moves back through time. For example, Eucommia tends to slot in the early part of the succession, with Carya, Pterocarya, and Tsuga at increasingly later phases (e.g. van der Wiel and Wijmstra 1987a, b; Ravazzi and Rossignol-Strick 1995; Leroy and Seret 1996). It is important to underline, however, that a taxon’s relative position within the course of the succession may vary through time as a result of climatic or historical factors, but also possibly because of evolutionary changes (e.g. West 1980). A recurrent successional pattern is also seen in the Late Pliocene interglacials in southern Italy, with Quercus and sclerophylls, followed

Cenozoic Climate and Vegetation Change

by Taxodiaceae and Cathaya, then Tsuga, Cedrus, Abies, and Picea (Combourieu-Nebout 1993). Indeed a modification of the idealized glacial–interglacial cycle has been proposed for this period with four phases: open vegetation (dry), deciduous/sclerophyllous woodland (warm but still dry), subtropical humid forest (warm and humid), and altitudinal coniferous forest (colder but still humid) (modified from Combourieu-Nebout 1993). Although broadly similar ecological-physiognomic assemblages recur in interglacials in the Mediterranean region, their qualitative and quantitative composition differed from one stage to the next. For example, high values of Carpinus characterized the later part of the Last Interglacial, but its importance was much restricted in the Holocene. Fagus, on the other hand, attained significant values in the Holocene, while during the Last Interglacial it was virtually absent. Abies was important during MIS 11. These differences in tree successions between interglacials may be the direct result of differences in the climatic signature of particular warm stages (e.g. Watts 1988; Huntley and Webb 1989). Bennett et al. (1991), however, suggested that the situation may be more complex because of interspecies and speciesclimate interactions adding an element of stochasticity in the character of the succession, while Tzedakis and Bennett (1995) drew attention to the importance of initial conditions and specifically to the nature of surviving populations during a preceding cold stage in influencing the course of interglacial succession. Tzedakis et al. (2001) explored the differences between individual temperate stages of the last 430 kyr by considering the records of several taxa, which form the main components of forest succession at the four longest pollen records from southern Europe. This revealed that Quercus, Olea, Ulmus, Tilia, and Corylus had a consistent pattern of presence from one stage to the next. By comparison Carpinus, Fagus, Abies, Buxus, and Pterocarya exhibited a variable behaviour in terms of presence/ absence, perhaps as a result of (1) their late expansion within the interglacial succession or (2) their reduced genetic variability. More specifically, it is possible that small variations in the pattern of vegetation succession, arising from slight differences in climate or historical accidents of founding and occurring early in the course of a temperate stage, could be magnified through time and influence the character of the ensuing succession and the chances for establishment and expansion of latecomers. In other words, the later a taxon normally expands, the higher the probability that it may not be able to attain its usual niche and therefore on separate occasions Carpinus, Fagus, Abies, Buxus, and Pterocarya may have been at a disadvantage. Alternatively, the

117

inconsistent presence of these taxa may be associated with their low species: genus ratio and reduced genetic variability in comparison with that of taxa such as Quercus and Ulmus. Adverse climatic conditions or the spread of disease during a temperate stage or the preceding cold stage have a higher probability of influencing taxa of reduced diversity. Climate pacing and the duration of interglacial vegetation succession. During the past 1 Myr interglacials had, on average, a duration of about half a precession cycle (e.g. Shackleton 1969). The timing of the early- and late-temperate part (i.e. zones II and III of the Turner and West (1968) scheme) of the interglacial succession is coincident with that part of the boreal summer insolation curve between the maximum and minimum points. More specifically, the interglacial forest expansion is closely associated with the timing of the summer insolation peak (e.g. Tzedakis 2005). Indeed, the floristic character of the succession, to a large extent, reflects the influence of insolation changes (e.g. Watts 1988), with mediterranean sclerophylls and other summerdrought resistant taxa expanding during the period of maximum summer insolation, while Fagus and Abies are better suited to the less-seasonal climates of the later part of interglacials. Figure 4.5 shows the main components of the forest succession during the Last Interglacial at Ioannina. The temperate phases containing the main part of the interglacial succession (Quercus, mediterranean taxa, Carpinus, Abies) occur between 127 and 115 ka. The post-temperate phase extends from 115 to 111 ka and is dominated by deciduous Quercus, but with forest becoming increasingly more open in character (Tzedakis et al. 2002a; 2003b). Deteriorating conditions eventually led to the disappearance of trees, although the exact timing may vary (see later sections). A clear precessional pattern is also seen during the interval preceding the onset of major NH glaciation. Forest periods alternate with open vegetation periods over the interval 3.0–2.92 Ma in a marine section from Sicily (Combourieu-Nebout et al. 2004), and a cyclic precessional pattern is also seen in the sections from the Ptolemais and Servia basins in northern Greece during the Mio-Pliocene (Kloosterboer-van Hoeve 2000). The question is how does the length of interglacial succession relate to changes in the dominant tempo of climate variability and more specifically, what happens during the interval of obliquity-dominated glacial cycles from 2.8 Ma to ∼1 Ma? Palynological analysis of an Early Pleistocene vegetational cycle (‘M’) in the Leffe Basin, northern Italy, has suggested a duration of approximately 30 kyr (Ravazzi and Rossignol-Strick 1995). The implication of this appears to be that during the Late

118

Chronis Tzedakis 3

d18 Obenthic (‰)

S06 3.5

4 4.5

5

Tenaghi Philippon

tree pollen (%)

100 75 50 25 0 600

June 65°N insolation Wm-2

550

500 450

400 1000

1100

1200

1300

Age (ka) Fig. 4.12. Variations in ‰18 O composition of benthic foraminifera in the S06 composite record (see the caption to Figure 4.7) over the interval 960–1,340 ka. Also shown are the June insolation for 65◦ N (Berger 1978) and total tree (dotted line) and temperate (total tree excluding Pinus and Juniperus (solid line)) pollen percentages at Tenaghi Philippon, north-eastern Greece (van der Wiel and Wijmstra 1987a, b). Tenaghi Philippon timescale is based on Tzedakis et al. (2006).

Pliocene/Early Pleistocene, interglacial vegetation succession in the Mediterranean may have been decoupled from the fundamental precessional pacing of Mediterranean climate (see earlier section on ‘Tempo of Mediterranean climate variability’), and instead become dominated by the high-latitude 41-kyr glacial cycle. However, closer examination of the Leffe sequence shows the presence of two vegetation successions within the interglacial part, the first characterized by expansion of Quercus, Ulmus, Carpinus, then Carya, Fraxinus, Pterocarya, Tsuga, closing with a large expansion of conifers (mainly Picea), that in turn is followed by a second broadly similar succession, with Quercus, Ulmus,

Carpinus, Carya, terminating in a prolonged coniferous phase, before the onset of a relatively short steppe vegetation phase. Indeed, in a recent re-evaluation of the chronological setting of the entire Leffe succession, the presence of two separate vegetational successions has now been recognized and they are designated as separate cycles ‘L’ and ‘M’ (Muttoni et al. 2007). Thus, the duration of the temperate part of an interglacial succession appears to be as long as in the Late Pleistocene, and approximately half a precessional cycle. This is further examined in Figure 4.12, which highlights the overall similarity between the ‰18 Obenthic record and tree pollen frequencies at Tenaghi

Cenozoic Climate and Vegetation Change

Philippon in terms of the obliquity-dominated glacial– interglacial stages, over the interval 0.96–1.35 Ma. However, the diagram also shows that the terrestrial record contains more internal structure, represented as distinct phases of vegetation development, than the ‰18 O record within individual stages. In fact the pollen record clearly shows oscillations in AP values that are closely following the precessional insolation changes and do not appear to be accompanied by major ice volume changes. Joint time-frequency (wavelet) analysis of the Tenaghi Philippon data supports the clustering of strong climatic precession cycles during intervals within the ‘41-kyr world’ (Tzedakis et al. 2006). This is echoed in recent spectral analyses of the palynological data from the Semaforo record (Calabria, Italy) over the interval 2.46–2.11 Ma, which show that vegetation varied at both the obliquity and precession bands (Klotz et al. 2006). Thus, the available evidence suggests that precession appears to be a fundamental pacing of the interglacial vegetation succession in the Mediterranean, with high-latitude effects superimposed after 2.8 Ma.

Meso-scale (102–3 yr) Glacial Sub-orbital Variability An important development in our understanding of Quaternary environments has been the realization of the pervasive and extreme nature of millennial- and centennial-scale climate variability, especially during intervals of increased ice volume. This has dispelled previous notions of glacial monotony and replaced them with a view of dynamic and unstable climate regimes. High-amplitude air-temperature decreases have been recognized in the ice-core records in Greenland (Greenland stadials or GS) and shown to be coeval with iceberg discharges and sea surface temperature (SST) reductions (the most extreme of which are known as Heinrich Events (HEs)) in the North Atlantic throughout the past 110 kyr (e.g. Dansgaard et al. 1993; Bond et al. 1993) and indeed during earlier glacials (e.g. Raymo et al. 1998; McManus et al. 1999). These changes are associated with disruptions of the meridional overturning circulation (MOC), with North Atlantic Deep Water (NADW) formation shifting from the Nordic seas to the subpolar North Atlantic during stadials, while during HEs, NADW was interrupted and the MOC was nearly, or completely eliminated (e.g. Rahmstorf 2002; McManus et al. 2004). These events were followed by a rapid resumption of the MOC and abrupt (within decades) warming in the order of

119

5–10◦ C, signalling the establishment of interstadial (GIS) conditions, sometimes lasting several millennia, before another rapid drop to stadial conditions. Spectral analyses have found an underlying periodicity of ∼1,500 yr in the glaciochemical series from the ice cores (e.g. Mayewski et al. 1997). Further statistical analyses have shown that the ‘waiting times’ between consecutive events are close to 1,500 yr, with additional preferred intervals at ∼3,000 and 4,500 yr, suggesting the operation of a 1,500-yr cycle, with sometimes a beat or two being skipped (Alley 2001; Rahmstorf 2003). This 1,500-yr cycle has been attributed to either periodic external (solar) forcing (e.g. Mayewski et al. 1997; van Geel et al. 1999), or to internal oscillations of the climate system (e.g. Broecker et al. 1990). However, the regularity of the cyclicity suggests that internal oscillations of the climate system, with its ubiquitous stochastic variability, are an unlikely explanation (Rahmstorf 2003). The extension of North Atlantic climate variability to the Mediterranean region has been associated with polar vortex expansion during HEs and GSs, leading to cooling and aridity by outbreaks of polar or continental air (e.g. Sánchez Goñi et al. 2002; Rohling et al. 2003). In addition, the higher atmospheric pressure gradient in the North Atlantic during these cold intervals led to more vigorous Saharan winds and an increased meridional dust transport to the Mediterranean (Moreno et al. 2005). The entrance of polar water through the Strait of Gibraltar during HEs contributed to the most extreme SST declines of the last glacial in the western Mediterranean (Cacho et al. 1999). These changes would be propagated eastwards, as reductions in SSTs would lower evaporation and therefore the moisture content of low-pressure systems moving across the Mediterranean, thus further intensifying the aridity. Indeed speleothem records from Israel provide clear evidence of cold and arid events associated with HEs (BarMatthews 1999). The record from Lake Lisan in the Dead Sea Transform, shows that superimposed on the longer-term trends of the high glacial lake levels were abrupt catastrophic drops in the order of 20–70 m, coeval with the timing of HEs (Bartov et al. 2004). Moreover, detailed examination of the varved sediments of Lake Lisan over the interval 17.7–26.2 ka, reveals that not only HEs, but also smaller ice-rafting events during intervening GSs were associated with arid intervals in the eastern Mediterranean (Prasad et al. 2004). In terms of the Mediterranean thermohaline circulation, low SSTs and cold winds during GSs led to enhanced deep water formation, while conversely warmer GIS conditions led to reduced deep water convection (Cacho

120

Chronis Tzedakis

et al. 2000; Sierro et al. 2005). However, during peak conditions of HEs the entry of polar water into the western Mediterranean led to reduced surface salinities and deep water formation (Sierro et al. 2005). In general, intervening interstadials (GIS) were characterized by higher sea surface and air temperatures, and increased moisture availability (Cacho et al. 1999; Sánchez Goñi et al. 2002; Moreno et al. 2005). That these climate changes must have had a major impact on vegetation across the Mediterranean is a reasonable enough assumption, but concrete evidence linking vegetation changes to North Atlantic events has been difficult to assemble. The main problem is that the rapid succession of climate oscillations, some of them lasting 1,000 years or less, means that most dating techniques fall short of the precision required to assign correct ages and attempt meaningful phase comparisons. Thus, high-frequency oscillations in tree populations have been known for some time in pollen sequences from southern Europe (e.g. Follieri et al. 1998), but despite the emergence of records supported by superior chronologies (e.g. Allen et al. 1999), dating uncertainties were still too large to establish the exact phase relationship with North Atlantic climate variability. Joint pollen and palaeoceanographic analyses in the same marine sequence have provided a solution to this problem by allowing in situ comparisons between pollen and proxy indicators of ocean environment. Of particular significance are results from deep ocean cores in the Portuguese margin and the Alboran Sea, western Mediterranean. These sequences have provided some of the most detailed marine oxygen isotope and SST records available for this interval (Figure 4.13), mirroring the Greenland ice core records event for event (e.g. Shackleton et al. 2000; Cacho et al. 1999). They have also furnished the first unequivocal evidence of the immediate response (within the sampling resolution of ∼200 years) of vegetation to sub-orbital-scale ´ et al. 2000, 2002; climate variability (Sánchez-Goni Roucoux et al. 2001, 2005; Combourieu Nebout et al. 2002). This work established that Iberian tree populations expanded and contracted repeatedly, closely tracking North Atlantic climate variability. Moreover, the Alboran Sea records showed a distinct separation in the magnitude of tree-population changes between HEs and intervening GSs (Sanchez-Goni 2002; Combourieu Nebout 2002), closely following the structure of SST changes in the same sequence (Figure 4.13). A similar pattern of large tree-population crashes during HEs and intermediate contractions during GSs can also be discerned in the pollen record from Lago Grande di Monticchio in Italy (Allen et al. 1999).

Tzedakis et al. (2004a) recently highlighted the importance of geographical position and local factors in producing a mosaic of patterns of sub-orbital-scale changes in climate and vegetation, by comparing three sites from contrasting environmental settings in Greece: (1) the Ioannina basin, an intramontane plateau on the western flank of the Pindus Mountains, characterized by high precipitation values and a sub-mediterranean climate (mean January temperature [Tjan ] 4.6◦ C; mean July temperature [Tjul ] 24.9◦ C; annual precipitation [Pann ] 1,200 mm); (2) Kopais, central Greece, located in the rain shadow of the Pindus, with a eu-mediterranean climate with little rainfall (mean Tjan 9◦ C; mean Tjul 27◦ C; Pann 470 mm); and (3) Tenaghi Philippon, north-east Greece, located in the landlocked Drama plain, with a distinctly more continental climate with occasional incursions of cold polar air (mean Tjan 3.4◦ C; mean Tjul 23.9◦ C; Pann 600 mm). Figure 4.13 shows the total percentages of temperate tree pollen from the Ioannina (Tzedakis et al. 2002b), Kopais (Tzedakis 1999), and Tenaghi Philippon (Wijmstra 1969) records for the interval 11–52 ka. The Tenaghi Philippon record is characterized by an almost complete absence of temperate tree populations both during HEs and intervening GSs, interspersed with low interstadial increases. The Kopais record also shows significant population crashes during HEs, while intervening GSs have intermediate declines. The Ioannina record shows large HE reductions and intermediate GS contractions, but the level of representation and continuity of pollen curves suggest that even the more extreme events were never severe enough to lead to the complete elimination of local populations. Moisture deficiency at both Kopais and Tenaghi Philippon would have led to significant reductions in temperate tree populations. Moreover, at Tenaghi Philippon, lower winter temperatures with frequent frosts would have had a further impact, leading to a complete absence of temperate trees during stadials. By comparison, at Ioannina continued moisture availability combined with high topographic variability would have allowed vertical migration of tree populations and generated a range of microhabitats suitable for survival. Thus, although climatic oscillations occurred in parallel at all three sites throughout the last glacial, their impact on tree populations was determined by the extent to which these changes crossed ecological thresholds. What emerges is that differences in the magnitude of HEs and GSs as recognized in the North Atlantic and western Mediterranean are expressed in terms of tree population changes in those areas with a range of favourable habitats, such as Ioannina. By contrast, the pollen record from Tenaghi Philippon, where glacial tree populations

Cenozoic Climate and Vegetation Change

121

_ _ _

Fig. 4.13. (a) Variations in ‰18 O composition of ice in the GISP2 (Greenland Ice Sheet Project 2) record (Stuiver and Grootes 2000). (b) Variations in ‰18 O composition of planktonic foraminifera in core MD95-2042 from the Portuguese margin (Shackleton et al. 2000). (c) Variations in alkenone-derived sea surface temperatures in marine core MD95-2043 from the Alboran Sea, western Mediterranean (Cacho et al. 1999). (d) Temperate tree pollen percentages in marine core MD95-2043 from the Alboran Sea (Sánchez Goñi et al. 2002). (e) Temperate tree pollen percentage curves from Ioannina I-284 (solid line), Kopais K93 (dotted line), and Tenaghi Philippon TF II (dashed line), Greece. Terrestrial age models according to Tzedakis et al. (2004a).

122

Chronis Tzedakis

were near their tolerance limits, do not appear to resolve differences in the amplitude of climate oscillations. This draws attention to the limits of the ‘sensitivity’ of different archives, which needs to be taken into account when mapping the spatial patterns of millennial climate variability (Tzedakis et al. 2004a). While evidence on the response of vegetation from the last glacial is slowly beginning to accumulate, information on earlier glacials in the Mediterranean is virtually non-existent. Comparison of changes in vegetation in north-east Greece over the last five glacial periods with the North Atlantic record of ODP 980 (McManus et al. 1999), showed similar repeat times in peaks of steppe vegetation with ice-rafting events, but the amplitude of these changes was not always proportional (Tzedakis et al. 2003c). However, this comparison was at the limit of the resolution of the terrestrial record to monitor high-frequency changes. New, highresolution records from marine sequences are needed to examine the nature of sub-orbital changes in climate and vegetation during earlier glacials. Given that the amplitude of this variability is likely to have been influenced by different boundary conditions that characterized previous glacials, it would be interesting to explore whether variations in the intensity of the forcing could lead to differences in the amplitude of the vegetation response.

Interglacial Sub-orbital Variability Holocene. Denton and Karlén (1973) were first to raise the possibility of Holocene climatic variations on the basis of synchronous advances in mountain glaciers in North America and Europe. Bond et al. (1997) showed that a pervasive 1,500-yr cycle in ice-rafted detritus in North Atlantic marine sediments of the last glacial extended into the Holocene. Although the amplitude of the Holocene changes was much subdued in relation to the last glacial, they nevertheless indicated a series of shifts in ocean surface hydrography associated with the southward advection of drift ice and cooler waters from the Nordic and Labrador Seas. A similar pacing in drift-ice proxies has also been found in Last Interglacial sediments in the North Atlantic, suggesting that a 1,500-yr cycle may have been a feature of past interglacial climates (Bond et al. 2001). Correlation of Holocene changes in drift ice indices with production rates of cosmogenic nuclides suggests that a solar forcing mechanism may underlie the 1,500yr cycle (van Geel et al. 1999; Bond et al. 2001). An additional ∼2,300-yr cycle of heliomagnetic origin is detected in the Holocene glaciochemical series from the

ice cores (Mayewski et al. 1997) which is also present in the ‰14 C residual series from tree rings (Stuiver and Braziunas 1989). A possible scenario linking variations in solar output with North Atlantic changes is that a decrease in UV radiation can lead to a reduction of stratospheric ozone, which in turn can cause high latitude cooling, an equatorward shift in subtropical jets, contraction of the Hadley cells and expansion of polar cells and the relocation of mid-latitude storm tracks (Haigh 1996; van Geel et al. 1999). In addition, the surface hydrographic changes may have led to reduced North Atlantic Deep Water formation, thereby amplifying the solar signal (Bond et al. 2001). However, the amplitude and geographical extent of these Holocene changes is limited, possibly because the ocean circulation is not close to a threshold situation during a warm climate (e.g. Ganopolski and Rahmstorf 2001). By comparison, during glacials climatic perturbations are significantly amplified through the presence of ice sheets and associated topographic and albedo effects, and threshold changes in iceberg discharges and ocean circulation (e.g. Shinn and Barron 1989; McManus et al. 1999; Ganopolski and Rahmstorf 2001). Most prominent of the Holocene events is an abrupt climate change around 8.2 ka reported not only from the immediate circum-North Atlantic region, but also from sites across the Northern Hemisphere (e.g. Alley and Ágústsdóttir 2005). The ‘8.2 ka event’ forms part of the series of cold Holocene events identified by Bond et al. (1997), but appears to have been amplified by the catastrophic drainage of glacial Lake Aggasiz, which freshened the North Atlantic (e.g. Clarke et al. 2004). Indeed, marine records of sufficient resolution show a short-lived, highamplitude anomaly embedded within a wider, loweramplitude cooling event (e.g. Alley and Ágústsdóttir 2005). Ice core records suggest that the climatic anomalies never reached a plateau of steady conditions and that the initial abrupt cooling took place within a decade and was followed by a period lasting up to 100–200 yr of return to conditions preceding the event (see ibid. and references therein). Given the subdued nature of Holocene climate cyclicity, its detection in Mediterranean proxy records has proven difficult. Alkenone-based SST reconstructions in the western Mediterranean show cooling oscillations occurring with a periodicity of 730 ± 40 yr or multiples thereof, with the amplitude of the cold events increasing towards the east, from 1.5◦ C in the Alboran to 3◦ C in the Tyrrhenian (Cacho et al. 2001). More recently, oxygen isotope analyses and abundances of planktonic foraminifera in the Tyrrhenian Sea have revealed the presence of at least seven cold episodes within the last

Cenozoic Climate and Vegetation Change

12 kyr (Sbaffi et al. 2004). By comparison, Rohling et al. (2002b) identified a series of cold events associated with northerly polar/continental air outbreaks over the Aegean Sea that occurred with a spacing of around 2,500 yr, i.e. closer to the 2,300-yr, rather than the 1,500-yr cycle. The SST changes based on planktonic foraminiferal abundances were in the order of 2–4◦ C. Turning to the terrestrial record, mineralogical, diatom, and ostracod evidence from Lake Tigalmamine in the Middle Atlas of Morocco show the occurrence of five century-scale arid Holocene intervals, reflecting reduced winter precipitation (Lamb et al. 1995). Of particular interest is that these well-defined arid events do not appear to have had any major impact on forest vegetation, indicating that precipitation remained sufficient for tree growth. Palaeolimnological data from Laguna de Medina, a saline lake in south-west Spain, show a series of abrupt desiccation events; the interval around 8 ka culminates in desiccation followed by a complete shift in ecology and is correlated with the 8.2 ka event, while correlation of the later desiccation events to North Atlantic changes is more tenuous (Reed et al. 2001). Sharp decreases in productivity, lasting 200–300 yr and associated with changes in temperature and/or effective moisture, are identified at 8.2, 6.4, and 3.8 ka in Lake Albano, central Italy (Ariztegui et al. 2001). In terms of the ‘8.2 event’, Mediterranean marine records of sufficient resolution show associated SST declines (e.g. Cacho et al. 2001; Rohling et al. 2002b), but the magnitude of the change is not significantly different from other Holocene cooling events. A prominent feature of Mediterranean marine sequences is an interruption in the deposition of sapropel S1, indicating a period of improved deep water ventilation between 8.5 and 8 ka (Chapter 2). This ventilation is associated with northerly cold outbursts over the eastern Mediterranean and probably related to the 8.2 ka event (Rohling et al. 2002b). During the interval of increased precipitation coeval with the deposition of S1, a distinct interruption is also seen in the speleothem record from Soreq Cave with ‰13 C values dropping between 8.2 and 8 ka, indicating a regional cooling and decrease in rainfall intensity (BarMatthews 1999). In terms of Holocene changes in vegetation, there is no immediately apparent evidence for a 1,500-yr or a 2,300-yr cycle. With regard to the more prominent 8.2 ka event, a forest regression episode is recorded in the west Cantabrian coastal mountains of north-west Iberia, which is under the direct influence of the North Atlantic (e.g. Muñoz Sobrino et al. 2005). Within the Mediterranean realm, a decrease in forest biomass is indicated by significant drops in pollen concentrations at Lago di Vico and Lagaccione, central Italy

123

(Magri and Sadori 1999; Magri 1999; Magri and Parra 2002) at ∼8.1 ka. However, a clear vegetation change in response to the 8.2 ka event is difficult to detect in most pollen sequences from the Mediterranean. In view of the short duration of the event, there are four, not mutually exclusive, reasons for this: (1) insufficient sampling resolution; (2) lack of pollen concentration data, which are better able to monitor changes in biomass than pollen percentages; (3) lack of chronological precision; (4) anthropogenic effects, masking natural changes. These problems notwithstanding, it is also possible that the magnitude of the event never exceeded the tolerance threshold for tree growth in areas of sufficient moisture availability. This, combined with the fact that most trees could outlive the duration of the event, may explain the paucity of evidence in the Mediterranean region. However, it is interesting to note that after 8.2 ka climatic conditions did not return to the early Holocene situation of increased summer aridity during the peak in boreal summer insolation. This change is chronologically best constrained in the annually laminated record of Lake Van, where oxygen isotopes and Mg/Ca ratios show the establishment of maximum Holocene lake levels and pollen data show the onset of Quercus expansion immediately after 8.2 ka (Wick et al. 2003). In Greece, maximum lake levels were reached around 8 ka, coincident with the onset of the Ostrya expansion (e.g. Digerfeldt et al. 2000). In the central and western Mediterranean, this interval appears to coincide with a decrease in Corylus and a rise in Quercus, Fagus, Abies, etc., depending on location. It is possible, therefore, that the initial climatic perturbation at 8.2 ka facilitated interspecific competition, leading to the establishment of different vegetation communities under a regime of reduced summer aridity. This has already been shown in annually laminated records from the Swiss plateau and southern Germany, where a sudden collapse of Corylus accompanied by a transient increase in Pinus, Betula, and Tilia coeval at 8.2 ka, was followed by the expansion of drought-sensitive species (Fagus) at 8.1 ka, outcompeting the others and forming a dense forest canopy (Tinner and Lotter 2001). While, early Holocene conditions with frequent summer droughts led to crown thinning and favoured the persistence of the light-demanding and short-grown Corylus, the 8.2 ka climatic oscillation followed by a shift towards reduced summer aridity led to Corylus being outcompeted by more shade-tolerant taxa (Tinner and Lotter 2001). Last Interglacial. Evidence of suborbital climate variability is available from earlier interglacials, but here the degree of chronological precision is increasingly limited

124

Chronis Tzedakis

and thus correlations become inevitably more tenuous. During the Last Interglacial, a series of coolings in the Nordic Seas have been detected around 127–126, 124, 122–121, and ∼117 ka (Fronval and Jansen 1997), while cold events in the Austrian Alps are indicated by ceased stalagmite growth between ∼130.7–125.7, ∼123.8, ∼120.5 ka, and after ∼118.2 ka (Holzkämper et al. 2004). A shift in oceanic conditions and a cold event is observed in the North Atlantic around 118 ka (e.g. Adkins et al. 1997; Lehman et al. 2002) and this may be coeval with a reduction in forest cover observed in French, Italian, and Greek sites (de Beaulieu and Reille 1992; Reille et al. 1998; Tzedakis et al. 2003b; Allen 2005). Joint pollen and stable isotope analyses on authigenic carbonates from the Last Interglacial at Ioannina, Greece, show isotopic shifts at approximately 127.3, 126.8, 125.7, 123.6, 120.3, and 114.2 ka, representing changes in the precipitation/evaporation ratio (Tzedakis et al. 2003b). What is remarkable is that every transition in the isotopic curves coincides with the onset of major changes in pollen assemblages, indicating the establishment of different vegetation communities. It is possible that these transitions are related to the North Atlantic cooling events identified by Bond et al. (2001) during the Last Interglacial. Although the magnitude of these events may not have been large enough to cause the demise of tree populations, it may have led to an increase in disturbance regimes that favoured the replacement of vegetation communities under new climatic conditions, reminiscent of some of the changes observed during the Holocene. Earlier interglacials. Further evidence concerning intra-interglacial variability has recently become available from joint foraminiferal oxygen isotope and pollen analyses in a deep-sea sequence off south-west Portugal (Tzedakis et al. 2004b). The record shows forest contractions within MIS 7e and 9e, which do not appear to have been triggered by changes in offshore conditions, but were coeval with sharp declines in atmospheric methane as recorded in Antarctic ice cores. Similar tree population reductions are observed in terrestrial pollen sequences across Mediterranean Europe (Reille and de Beaulieu 1995; Reille et al. 1998; Follieri et al. 1988; Wijmstra and Smit 1976), pointing to the wide geographical extent of these changes. The confluence of atmospheric methane and vegetation changes draws attention to the occurrence of substantial (rather than subdued) events occurring on a global scale during intervals of low ice volume. Given that most prominent abrupt climate events have hitherto been usually associated with glacial climates, this evidence underscores the importance of understanding the origin

of this intra-interglacial variability (Tzedakis et al. 2004b).

Interactions Between Scales of Variability The discussion presented thus far has been organized under the distinct headings representing different scales of environmental variability. However, this artificial separation is missing an important element of the story in terms of the interactions between the different scales.

Macro- and Meso-scale Interactions Tzedakis (2005) reviewed the interaction between orbital and sub-orbital variability and found the following patterns in terms of the response of vegetation in southern Europe. 1. Duration of forest periods. In the closing stages of interglacials, forest decline in southern Europe may often lag glacial inception by a few millennia, but abrupt climate events of sufficient magnitude can lead to an early demise of tree populations and a premature ending of the forest period. Thus, the end of the Last Interglacial forest period occurred 5 ka after glacial inception (e.g. Shackleton et al. 2002, 2003) and was caused by an ice-rafting event, disrupting the MOC and leading to a reduction of moisture availability in southern Europe. Similar mechanisms led to the end of the two ensuing forest intervals, within MIS 5. The difference, however, is that while the ice-rafting event in the first case occurred ∼6 kyr after the glacial inception, much in the same fashion as in the Last Interglacial, the second occurred much earlier, right at the MIS 5a/4 boundary and resulted in a considerably shorter forest period (c.7 kyr) compared to the two earlier periods (∼16 kyr). Recent results from joint pollen and marine proxy analyses in a deep-sea sequence on the Portuguese margin, discussed in the previous section, also showed that the duration of forest periods over the last 350 kyr varied considerably (Tzedakis et al. 2004b). In this case, forest contractions do not seem to be associated with North Atlantic ice-rafting events, but were coeval with abrupt declines in atmospheric methane, after which tree populations in western Iberia, but also France, Italy, and Greece, did not always recover, leading to a premature ending of the forest stages. What emerges is that although the broad timing of interglacials is consistent with orbital (Milankovitch) theory, their specific

Cenozoic Climate and Vegetation Change

duration appears to be dictated by sub-orbital climate variability rather than the length of the half-precession cycle (Tzedakis et al. 2004b). 2. Amplitude of sub-orbital-scale changes. Although southern European pollen records show a close correspondence with North Atlantic variability in terms of the number of events, they may diverge in terms of the amplitude. These differences in amplitude should be placed within the context of orbital mean state as reflected in long-term insolation and ice volume changes. Thus, the presence of significant ice sheets in MIS 2 appears to have been sufficient to suppress changes in temperate tree population size in response to North Atlantic variability. Conversely, the higher insolation and smaller ice volume of the early MIS 3 interval must have contributed to the relative stability of the period (i.e. reduced occurrence of ice-rafting events) and by extension to the prolonged duration of GIS 14 and GIS 16. In addition, it allowed large tree population expansions in the Mediterranean as suggested by high pollen percentages and concentrations. These cases underscore the importance of orbital mean state in modulating the amplitude of the vegetation response to sub-orbital-scale oscillations (Tzedakis 2005).

Mega-, Macro- and Meso-scale Interactions Climate variability and evolutionary responses. In contrast to earlier refuge hypotheses that population fragmentation during glacial stages of the Quaternary led to the intensification of speciation (Haffer 1969), a more recent view is that the accentuated environmental instability did not lead to increased speciation rates, with most species predating the Pleistocene (e.g. Bennett 1997; Klicka and Zink 1997). This seeming evolutionary stability has been attributed to orbital climate fluctuations, which undo microevolutionary changes by forcing repeated population crashes, range shifts, and gene flow (Bennett 1997; Dynesius and Jansson 2000). Central to this view is that the alternation of climate phases does not generally allow sufficient time for long-term isolation and genetic differentiation to species rank, or drives incipient species to extinction. Moreover, the extreme nature of suborbital climate variability, especially during HEs, suggests that the waiting times between climatic perturbations were even shorter than that allowed by orbital changes (e.g. Roy et al. 1996). However, molecular genetic data reveal considerable divergence in southern European refugial centres, which took several glacial–interglacial cycles to accumulate (Hewitt 1996, 1999, 2000). More general

125

analysis of DNA divergence in animals shows that species have formed through the Pliocene and Pleistocene and that such divergence has proceeded apparently unhindered in some places (Hewitt 1996; Avise et al. 1998). Thus, while in lowland tropical forests most species of birds formed before the Quaternary, clusters of recently diverged lineages along with older species are found in tropical mountain regions (Fjeldså and Lovett 1997). These mountains are considered centres for speciation because they provide a relatively stable habitat through climate oscillations in which older species survive and new lineages are generated. Although no direct palaeoenvironmental evidence has been available, this long-term stability is thought to be a function of continued moisture availability and varied topography (ibid.; Hewitt 2000). Thus, a revised view of Quaternary evolutionary trends is that while climate variability mostly inhibited speciation, species continued to form in places where presumed ecological stability allowed the accumulation of genetic divergence over several glacial– interglacial cycles (Tzedakis et al. 2002b). Recent palaeoecological data suggest that midaltitude refugial sites in the Mediterranean may provide a mid-latitude analogue to the tropical mountain speciation centres of Fjeldså and Lovett (1997). More specifically, a high-resolution pollen record from the Ioannina basin, Greece, shows that local temperate tree populations managed to survive the impact of both orbital and millennial-scale climatic extremes (Tzedakis et al. 2002b). This provides evidence for the existence of an area of relative ecological stability, where local conditions (sustained moisture availability and varied topography) appear to have buffered the extreme effects of Quaternary climate variability, contributing to the survival of residual tree populations. When combined with an earlier pollen record from the same basin, spanning multiple glacial–interglacial cycles (Tzedakis 1993), what emerges is that populations of many temperate tree species have persisted in this general area, over several hundred thousand years, albeit at varying abundances. The richness of the Mediterranean flora with its unusually high endemism is, in part, a reflection of its geographical position and geological history, as well as the extent to which Tertiary species managed to survive the effects of Quaternary climate variability (Blondel and Aronson 1999; Chapter 23). However, local buffering from extreme environmental effects, as illustrated by the Ioannina record, not only led to reduced extinction rates but, when combined with genetic isolation, may have allowed lineage divergence to proceed through the Quaternary. Such ecologically stable areas may be critical

126

Chronis Tzedakis

not only for the long-term survival of species, but also for the emergence of new ones (Tzedakis et al. 2002b). The role of such areas in southern Europe has been considered by Hampe and Petit (2005), who underscored the critical importance of populations at the low-latitude limit of species ranges. In contrast to the ‘centre–periphery hypothesis’, which predicts that marginal populations are genetically less diverse, phylogeographic surveys show that as a consequence of climatedriven range modifications, it is marginal populations that commonly contain the majority of species’ genetic diversity (e.g. Hewitt 2000; Petit et al. 2003). Hampe and Petit (2005) point out that while much attention has been focused on processes at the leading edge of range expansions, there has been limited work on rear range margins and more specifically on what they term ‘stable rear edges’, where at least some populations have persisted through multiple Quaternary climate oscillations. These populations are characterized by local adaptation and genetic drift, resulting in distinct ecotypes and high levels of regional genetic diversity (Hampe and Petit 2005, and references therein). Given that most northern European populations are eliminated during glacials, Mediterranean refugia representing stable rear edge populations may be seen as both long-term stores of genetic diversity and also crucibles of speciation (e.g. Blondel and Aronson 1999; Tzedakis et al. 2002b, 2003a; Hampe and Petit 2005). Identification of such populations should be a research priority, but appropriate conservation practices need to take into account the peculiarity of rear edge populations, which are different from populations in other parts of the range (Hampe and Petit 2005).

Mega- and Macro-scale Interactions Origins of mediterranean climate and sclerophylly. The present-day broadleaved evergreen sclerophyll elements that dominate the vegetation in areas with mediterranean climate are considered to have their origins in the Tertiary laurophyll vegetation (Mai 1989). For example, evergreen Quercus, Olea, Pistacia, Phillyrea have their first appearance in the Oligocene and Miocene fossil record (Palamarev 1989). By comparison, the origin of a mediterranean-type climate is considered to have occurred much later, towards the end of the Tertiary, as a result of the intensification of NH glaciation and the development of steep thermal gradients and major cold currents (Axelrod 1973; Raven 1973). Simply put, the occurrence of the summer-dry, winterwet climate rhythm in the Mediterranean today is

controlled by the positioning and seasonal movement of the subtropical anticyclone belt on the equatorial side and the location of areas of cyclogenesis in the belt of mid-latitude westerlies on the poleward side (e.g. Deacon 1983; Chapter 3). Flohn (1978) has suggested that during the early Tertiary, under ice-free conditions and a low pole-to-equator temperature gradient, the subtropical anticyclone belt would have been located approximately between 50◦ and 60◦ latitude, with a seasonally shifting Hadley circulation bringing convective summer rainfall over 70–80 per cent of the global surface. After the glaciation of Antarctica in the Oligocene, there would have been an asymmetry in global climate, which was not restored until the intensification of NH glaciation (Flohn 1978). This would have moved the subtropical high-pressure cell into its present position and led to the establishment of mediterranean-type climates. Indeed, palynological evidence from the northwest Mediterranean sector has shown a major reduction in subtropical taxa and their replacement by sclerophyllous vegetation around 3.2 Ma (subsequently revised to 3.6 Ma) (Suc 1984; Suc and Popescu 2005) and this is the generally accepted age for the establishment of mediterranean-type climatic rhythm in the Mediterranean region. The above considerations led Axelrod (1973: 273) to state that ‘since the sclerophylls are older than the mediterranean climate in which they have survived, they must have been preadapted to summer drought’. According to Axelrod (1975), since the adaptive structural features of broadleaved evergreen sclerophylls (evergreenness, thick small leaves, sprouting habit, deep root system) are also found in mesic laurophyllous species, they could not have originated in response to summer drought. Other features (e.g. sunken stomata, reduced leaf size) may represent adaptations to drier conditions (but mainly winter aridity), which took place in rain-shadow areas, mineral deficient sites, etc. Thus, as summer rainfall decreased during the later Tertiary, these plants were preadapted to survive the mediterranean-type summer aridity (Axelrod 1975). Sclerophylly in Axelrod’s view is an ancient and possibly primitive feature and from an evolutionary standpoint he considers sclerophylls as ‘generalists’ living in diverse environments and able to adjust to environmental changes. Over the last quarter of a century the idea that the evolution of sclerophylly significantly predates the appearance of mediterranean-type climates has become part of the current orthodoxy and mainstream literature. However, some of the basic premises regarding the evolution of mediterranean climates may not stand up to close

Cenozoic Climate and Vegetation Change

scrutiny. Thus, the poleward extension of the subtropical anticyclone belt to 50–60◦ latitude is not a feature of Cretaceous and early Tertiary climate simulations (Valdes, pers. comm.). Moreover, the latitudinal distribution of evaporites, phosphorites, and coal deposits during the Maastrichtian (65–71 Ma) indicates that the zone of subtropical high pressure was centred near 25– 30◦ palaeolatitude, suggesting that the basic components of the present-day Hadley circulation were also active during the Late Cretaceous (Horrell 1991). On the basis of the distribution of these geological deposits and fossil plants, Horrell (1991) reconstructed an extensive winter-wet biome on the poleward side of subtropical high belt. There is also evidence that a mediterraneantype climatic rhythm may have been present in the Late Jurassic at about 36◦ palaeolatitude, from fossils of Cheirolepidaceae bearing tree-ring patterns similar to those of modern conifers from the Mediterranean (including ‘false’ or ‘intra-annual’ rings), which suggests water shortage during the growing season (Francis 1984). These finds are associated with evaporites and calcrete crusts in palaeosols, which are also indicative of mediterranean conditions (ibid.). In the Neogene, examination of palaeosols in Greece and Turkey over the last 11 Myr showed the presence of abundant soil carbonate, suggesting a rainfall regime of less than 1,000 mm yr−1 (Quade et al. 1994). Moreover, carbon isotopic values from the palaeosols and fossil teeth showed that vegetation was dominated by C3 plants, while C4 plants were missing, suggesting that regimes with high summer rainfall can be excluded (Quade et al. 1994). Given that with sufficient resources C3 trees will outcompete C4 grasses (Collatz et al. 1998), the absence of C4 grasses in the Miocene may not necessarily be a reflection of climatic conditions. However, pollen spectra from the eastern Mediterranean show the presence of grasses (∼20%) in the Miocene, which are deemed to be C3 on the basis of the isotopic evidence (e.g. Bertolani Marchetti and Accorsi 1978). This suggests that the exclusion of C4 grasses was climatically controlled and therefore points to the occurrence of regimes with reduced summer precipitation back to ∼11 Ma. One feature that is consistently missing from the above narrative is the effect of orbital variations. As discussed earlier, precession has been a fundamental pacing of climate in the Mediterranean region. At times of boreal insolation maxima this would lead to the intensification of summer aridity and also increased winter precipitation. While this essentially corresponds to an accentuated mediterranean climate rhythm, its effects would have been limited during the Tertiary by ample precipitation originating from the warm Tethyan and

127

Mediterranean Seas. However, if moisture availability became relatively limited at particular locations or during specific geological intervals, the effects of orbital cyclicity would have been more pronounced. This can be seen in the palynological results from the Ptolemais and Servia basins in northern Greece, spanning three phases: before, during, and after the Messinian Salinity Crisis (Kloosterboer-van Hoeve 2000). The lithostratigraphic units of all three phases exhibit a distinctive precessional cyclicity, with regularly alternating organicrich and carbonate-rich beds. A duration of ∼21 kyr per sedimentary couplet has been determined through magnetostratigraphy and 40 Ar/39 Ar dating (Steenbrink et al. 1999). In the pre-MSC phase, vegetation was characterized by a mixture of deciduous and coniferous trees with subdued vegetation cyclicity. During the MSC, the record is strikingly different, showing pronounced cyclic changes between periods of herbaceous dominance and coniferous dominance. Of particular interest is the appearance of evergreen Quercus (previously missing) during transitional intervals. In the post-MSC phase, there was a cyclic succession with intervals of evergreen Quercus followed by deciduous Quercus and then Pinus. What emerges is that while precession-driven changes occurred throughout the studied interval, their effect on vegetation was modulated by the climate mean state. In the pre-MSC phase, changes in moisture levels were not extreme enough to limit tree growth or provide a competitive advantage for sclerophyllous taxa. However, during the MSC a substantially drier mean climate state meant that tolerance thresholds for most temperate tree growth were periodically exceeded, while sclerophylls had had intervals when they were climatically favoured. After the MSC, although moisture availability increased, periodic increases in evergreen Quercus and continued presence of herbs suggest that the bioclimatic situation was more susceptible to precessional variations compared to the pre-MSC phase. This suggests that conditions of increased seasonality characterized by a summer-dry, winter-wet climatic rhythm may have appeared intermittently during the course of the Tertiary, or indeed even before. This view of the ephemeral nature of mediterranean climates applies also to the Plio-Pleistocene, where the establishment of the mediterranean climatic rhythm around 3.6 Ma should not be confused with permanence. Mediterranean conditions would appear during interglacials (reaching their maximum expression during boreal summer insolation maxima), but would not persist during glacials. Indeed, ‘false’ tree rings appear to be absent from Mediterranean conifers of Late Glacial interstadial age (Friedrich, pers. comm.).

128

Chronis Tzedakis

The above discussion suggests that the mediterranean-type climate rhythm may have appeared intermittently during the course of the Tertiary (or indeed before), much earlier than hitherto considered. If that is correct, then the paradigm that sclerophylls were preadapted to summer drought because the mediterranean climate is a recent phenomenon, needs to be reconsidered.

Concluding Remarks The chapter traced the history of climate and vegetation changes in the Mediterranean over the last 65 Myr by extracting broad temporal trends and geographical patterns. While, wherever possible, I have tried to highlight the importance of local factors in shaping environmental responses within a hugely heterogeneous Mediterranean backdrop, inevitably the need to provide a structured sense of geographical and temporal changes may have led to some oversimplification. However, a central feature emerging from this review is that studies with improved resolution and chronological precision have invariably shown geographically coherent changes, where previously discordant patterns had been invoked. Although the amplitude of changes may differ between locations, the direction of change usually appears to be similar. The main conclusions of this review have been presented under each section. One final point that deserves some attention is the need for improved resolution and chronological precision. While the issue of resolution has been largely addressed in the upper part of the record, it is important that increasingly detailed studies (i.e. sampling resolution better than 200 yr) are undertaken in the Tertiary. As regards chronologies, the fact remains that uncertainties associated with virtually all dating methods preclude any detailed phase comparisons between different components of the climate system and geographical areas, and this inevitably deteriorates as one moves back through time. Efforts should focus on alternative methods aiming to synchronize separate records either by placing them on a common timescale or by linking them via the presence of markers such as volcanic tephras. However, the timing of volcanic eruptions and climatic transitions of particular interest may not necessarily coincide. Joint pollen and palaeoceanographic analyses within the same marine sequence appear at present to be the only method of allowing a stratigraphically continuous monitoring of phase relationships. In this respect, the emergence of an astronomically calibrated framework using the occur-

rence of sapropels and carbonate cycles for the last 12 Myr (e.g. Lourens et al. 2004), may provide major opportunities for significantly extending the terrestrial record back in time.

Acknowledgements I am indebted to Jamie Woodward for his ‘Byzantine’ invitation to contribute to this volume, and for his patience and support during the long gestation period. I am grateful to Gunnar Digerfeldt and Lucas Lourens for advice and comments on part of the manuscript and Jacques-Louis de Beaulieu and Donatella Magri for most helpful reviews. I thank Ian Lawson and Katy Roucoux for discussions and comments and Nick Shackleton for providing the new S06 composite benthic isotope record.

References Adkins, J. F., Boyle, E. A., Keigwin, L., and Cortijo, E. (1997), Variability of the North Atlantic thermohaline circulation during the last interglacial period. Nature 390: 154–6. Agwu, C. O. C. and Beug, H.-J. (1984), Palinologische Untersuchungen an marinen Sedimenten vor der westafrikanische Küste. Palaeoecology of Africa 16: 37–52. Allen, J. R. M., Brandt, U., Brauer, A., Hubberten, H.-W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J. F. W., Nowaczyk, N. R., Oberhänsli, H., Watts, W. A., Wulf, S., and Zolitschka, B. (1999), Rapid environmental changes in southern Europe during the last glacial period. Nature 400: 740–3. Watts, W. A., and Huntley, B. (2000), Weichselian palynostratigraphy, palaeovegetation and palaeoenvironments; the record from Lago Grande di Monticchio, southern Italy. Quaternary International 73/4: 91 9110. Alley, R. B. and Ágústsdóttir, A. M. (2005), The 8k event: cause and consequence of major Holocene abrupt climate change. Quaternary Science Reviews 24: 1123–49. Anandakrishnan, S., and Jung, P. (2001), Stochastic resonance in the North Atlantic. Paleoceanography 16: 190–8. Andersen, S. T. (1966), Interglacial vegetational succession and lake development in Denmark. Palaeobotanist 15: 117–27. Ariztegui, D., Chrondrogianni, C., Lami, A., Guilizzoni, P., and Lafargue, E. (2001), Lacustrine organic matter and the Holocene paleoenvironmental record of Lake Albano (central Italy). Journal of Paleolimnology 26: 283–92. Arz, H.W., Lamy, F., Pätzold, J., Müller, P. J., and Prins, M. (2003), Mediterranean moisture source for an earlyHolocene humid period in the northern Red Sea. Science 300: 118–21. Avise, J. C., Walker, D., and Johns, G. C. (1998), Speciation durations and Pleistocene effects on vertebrate phylogeography. Proceedings of the Royal Society LondonB 265: 1707–12. Axelrod, D. I. (1973), History of the Mediterranean ecosystem in California, in F. di Castri and H. A. Mooney (eds.), Mediterranean Type Ecosystems. Chapman & Hall, London, 225–77.

Cenozoic Climate and Vegetation Change (1975), Evolution and biogeography of Madrean-Tethyan sclerophyll vegetation. Annals of the Missouri Botanical Garden 62: 280–334. Ayalon, A., Bar-Matthews, M., and Kaufman, A. (1999), Petrography, strontium, barium and uranium concentrations, and strontium and uranium isotope ratios inspeleothems as palaeoclimatic proxies: Soreq Cave, Israel. The Holocene, 9: 715–22. Azzaroli, A. and Guazzone, G. (1979/80), Terrestrial mammals and land connections in the Mediterranean before and during the Messinian. Palaeogeograph, Palaeoclimatology, Palaeoecology 29: 155–67. Bar-Matthews, M., Ayalon, A., and Kaufman, A. (1997), Late Quaternary paleoclimate in the eastern Mediterranean region from stable isotope analysis of speleothems at Soreq Cave, Israel. Quaternary Research 47: 155–68. and Wasserburg, G. J. (1999), The Eastern Mediterranean paleoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166: 85–95. (2000), Timing and hydrological conditions of Sapropel events in the eastern Mediterranean, as evident from speleothems, Soreq cave, Israel. Chemical Geology 169: 145–56. Bard, E., Delaygue, G., Rostek, F., Antonioli, F., Silenzi, S., and Schrag, D. (2002), Hydrological conditions over the western Mediterranean basin during the deposition of the cold Sapropel 6 (ca. 175 kyr BP). Earth Planetary Science Letters 202: 481–94. Barron, E. J. and Pollard, D. (2002), High-resolution climate simulations of Oxygen Isotope Stage 3 in Europe. Quaternary Research 58: 296–309. van Andel, T. H., and Pollard, D. (2004), Glacial environments II: reconstructing the climate of Europe in the last Glaciation, in T. H. van Andel and S. W. Davies (eds.), Neanderthals and Modern Humans in the European Landscapes during the Last Glaciation. McDonald Institute for Archaeological Research, Cambridge, 57–78. Bartoli, G., Sarnthein, M., Weinelt, M., Erlenkeuser, H., GarbeSchönberg, D., and Lea, D. W. (2005), Final closure of Panama and the onset of northern hemisphere glaciation. Earth and Planetary Science Letters 237: 33–44. Bartov, U., Stein, M., Enzel, Y., Agnon, A., and Reches, Z. (2002), Lake levels and sequence stratigraphy of Lake Lisan, the Late Pleistocene precursor of the Dead Sea. Quaternary Research 57: 9–21. Goldstein, S. L., Stein, M., and Enzel, Y. (2004), Catastrophic arid episodes in the Eastern Mediterranean linked with the North Atlantic Heinrich events. Geology 31: 439–42. Baruch, U. and Bottema, S. (1999), A new pollen diagram from Lake Hula, in H. Kawagoe, G. W. Coulter, and A. C. Roosevelt (eds.), Ancient Lakes: Their Cultural and Biological Diversity. Kenobi Production, Ghent, 75–86. Begin, Z. B., Ehrlich, A., and Nathan, Y. (1974), Lake Lisan, the Pleistocene precursor of the Dead Sea. Geological Survey of Israel Bulletin 63. Ben Tiba, B. and Reille, M. (1982), Recherches pollenanalytiques dans les montagnes de Kroumirie (Tunisie septentrionale): premiers résultats. Ecologia Mediterranea 7: 75–86. Bennett, K. D. (1997), Evolution and Ecology, The Pace of Life. Cambridge University Press, Cambridge. Tzedakis, P. C., and Willis, K. J. (1991), Cold stage refugia of north European trees. Journal of Biogeography 18: 103–15.

129

Berger, A. (1978), Long-term variations of caloric insolation resulting from the earth’s orbital elements. Quaternary Research 9: 139–67. and Loutre, M. F. (1991), Insolation values for the climate of the last 10 million years. Quaternary Science Reviews 10: 297–317. and Wefer, G. (2003), On the dynamics of the Ice Ages: Stage-11 paradox, Mid-Brunhes climate shift, and 100-ky cycle. In A. Droxler (ed.), Earth’s Climate and Orbital Eccentricity: The Marine Isotope Stage 11 Question. Geophysical Monograph 137, American Geophysical Union, Washington, 41–59. Bertolani-Marchetti, D. (1985), Pollen paleoclimatology in the Mediterranean since Messinian time in D. J. Stanley and F. C. Wezel (eds.), Geological Evolution of the Mediterranean Basin. Springer, New York, 525–43. and Accorsi, C. A. (1978), Palynological studies on samples from DSDP Leg 42A, in K. J. Hsü, et al. (eds.), Initial Reports of the Deep Sea Drilling Project 42: 789–803. Bertoldi, R., Rio, D., and Thunnell, R. (1989), PliocenePleistocene vegetational and climatic evolution of the southcentral Mediterranean. Palaeogeograph, Palaeoclimatology, Palaeoecology 72: 263–75. Beug, H.-J. (1968), Probleme der Vegetationgeschichte in Südeuropa. Berichte der Deutschen Botanischen Gesellschaft 80: 682–9. (1975), Changes of climate and vegetation belts in the mountains of Mediterranean Europe during the Holocene. Biuletyn Geologiczny Warszawa 19: 101–10. Birks, H. J. B. (1986), Late-Quaternary biotic changes in terrestrial and lacustrine environments, with particular reference to north-west Europe, in B. E. Berglund (ed.), Handbook of Holocene Palaeoecology and Palaeohydrology, Wiley, Chichester, 3–65. Birks, H. J. B. and Line, J. M. (1993), Glacial refugia of European trees—a matter of chance? Dissertationes Botanicae 196: 283–91. Blondel, J. and Aronson, J. (1999), Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford. Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J., and Bonani, G. (1993), Correlations between climate records from North-Atlantic sediments and Greenland ice. Nature 365: 143–7. Showers, W., Cheseby, M., Lotti, R., Almasi, P., deMenocal, P., Priore, P., Cullen, H., Hajdas, I., and Bonani, G. (1997), A pervasive millennial-scale cycle in North Atlantic Holocene and glacial climates. Science 278: 1257–66. Kromer, B., Beer, J., Muscheler, R., Evans, M. N., Showers, W., Hoffmann, S., Lotti-Bond, R., Hajdas, I., and Bonani, G. (2001), Persistent solar influence on North Atlantic climate during the Holocene, Science 294: 2130–6. Bond, W. J., Woodward, F. I., and Midgley, G. F. (2005), The global distribution of ecosystems in a world without fire. New Phytologist 165: 525–38. Bottema, S. (1974), Late Quaternary vegetation history of Northwestern Greece, Ph.D. Thesis, Rijksuniversiteit te Groningen, Groningen. (1979), Pollen analytical investigations in Thessaly (Greece). Palaeohistoria 21: 19–40. and Sarpaki, A. (2003), Environmental change in Crete: a 9000-year record of Holocene vegetation history and the effect of the Santorini eruption. The Holocene 13: 733–49. Brauer, A., Allen, J. R. M., Mingram, J., Dulski, P., Wulff, S., and Huntley, B. (2007), Evidence for last interglacial chronology

130

Chronis Tzedakis

and environmental change from southern Europe. Proceedings of the National Academy of Sciences 104: 450–4. Bredenkamp, G. J., Spada, F., and Kazmierczak, E. (2002), On the origin of northern and southern hemisphere grasslands. Plant Ecology 163: 209–29. Brewer, S., Cheddadi, R., de Beaulieu, J. L., Reille, M., Data contributors (2002), The spread of deciduous Quercus throughout Europe since the last glacial period. Forest Ecology and Management 156: 27–48. Broecker, W. S., Bond, G., Klas, M., Bonani, G., and Wolfi, W. (1990), A salt oscillator in the glacial North Atlantic? 1. The concept. Paleoceanography 5: 469–77. Butzer, K. W. (1957), Mediterranean pluvials and the general circulation of the Pleistocene. Geografiska Annaler 39: 48–53. Cacho, I., Grimalt, J. O., Pelejero, C., Canals, M., Sierro, F. J., Flores, J. A., and Shackleton, N. (1999), Dansgaard-Oeschger and Heinrich event imprints in Alboran Sea paleotemperatures. Paleoceanography 14: 698–705. Sierro, F. J., Flores, J. A., Shackleton, N. J., and Canals, M. (2000), Evidence for enhanced Mediterranean thermohaline circulation during rapid climate coolings. Earth and Planetary Science Letters 183: 417–29. Canals, M., Sbaffi, L., Shackleton, N. J., Schönfeld, J., and Zahn, R. (2001), Variability of the western Mediterranean Sea surface temperature during the last 25,000 years and its connection with the Northern Hemisphere climatic changes. Paleoceanography 16: 40–52. Capraro, L. (2002), Riconstruzione della storia della vegetazione durante la transizione climatica del Pleistocene Medio (MPT) nel Mediterraneo Centrale. Tesi di Dottorato di Ricerca, Università degli Studi di Padova. Carrión, J. S. (2002), Patterns and processes of Late Quaternary environmental change in a montane region of southwestern Europe. Quaternary Science Reviews 21: 2047–66. Cerling, T. E., Harris, J. M., MacFadden, B. J., Leakey, M. G., Quade, J., Eisenmann, V., and Ehleringer, J. R. (1997), Global vegetation change through the Miocene/Pliocene boundary. Nature 389: 153–8. Cheddadi, R. and Rossignol-Strick, M. (1995), Eastern Mediterranean Quaternary paleoclimates from pollen and isotope records of marine cores in the Nile cone area. Paleoceanography 10: 291–300. Clarke, G. K. C., Leverington, D. W., Teller, J. T., Dyke, A. S. (2004), Paleohydraulics of the last outburst flood from glacial lake Agassiz and the 8200 BP cold event. Quaternary Science Reviews 23: 389–407. COHMAP Members (1988), Climatic changes of the last 18,000 years: observations and model simulations. Science, 241: 1043–52. Collatz, G., Berry, J. A., and Clark, J. S. (1998), Effects of climate and atmospheric CO2 partial pressure on the global distribution of C4 grasses: present, past, and future. Oecologia 114: 441–54. Combourieu-Nebout, N. (1993), Vegetation response to Upper Pliocene glacial/interglacial cyclicity in the central Mediterranean. Quaternary Research 40: 228–36. and Vergnaud-Grazzini, C. (1991), Late Pliocene Northern Hemisphere glaciations: the continental and marine responses in the central Mediterranean. Quaternary Science Reviews 10: 319–34. Turon, J. L., Zahn, R., Capotondi, L., Londeix, L., and Pahnke, K. (2002), Enhanced aridity and atmospheric high-

pressure stability over the western Mediterranean during the North Atlantic cold events of the past 50 k.y. Geology 30: 863–6. Foucault, A., and Mélières, F. (2004), Vegetation markers of palaeoclimate cyclical changes in the Pliocene of Punta Piccola (Sicily, Italy). Palaeogeography, Palaeoclimatology, Palaeoecology 214: 55–66. Cramer, B. S. (2001), Latest Paleocene-earliest Eocene cyclostratigraphy: using core photographs for reconnaissance geophysical logging. Earth and Planetary Science Letters 186: 231–44. Cowling, S. A. and Sykes, M. T. (1999), Physiological significance of low atmospheric CO2 for plant-climate interactions. Quaternary Research, 52: 237–42. de Beaulieu, J.-L. and Reille, M. (1992), Long Pleistocene pollen sequences from the Velay Plateau (Massif Central, France), I. Ribains maar. Vegetation History and Archaeobotany 1: 233–42. de Jong, J. (1988), Climatic variability during the past three million years, as indicated by vegetational evolution in northwest Europe and with emphasis on data from The Netherlands. Philosophical Transactions of the Royal Society London B 318: 603–17. deMenocal, P. B. (2004), African climate change and faunal evolution during the Pliocene-Pleistocene. Earth and Planetary Science Letters 220: 3–24. Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahljensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjornsdottir, A. E., Jouzel, J., and Bond, G. (1993), Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364: 218–20. Deacon, H. J. (1983), The comparative evolution of Mediterranean-type ecosystems: a southern perspective, in F. J. Kruger, D. T. Mitchell, and J. U. M. Jarvis (eds.), Mediterranean Type Ecosystems; The Role of Nutrients. Springer, Berlin, 3–40. Demesure, B., Comps, B., and Petit, R. J. (1996), Chloroplast DNA phylogeography of the common beech (Fagus sylvatica L.) in Europe. Evolution 50: 2515–20. Denton, G. H. and Karlén, W. (1973), Holocene climatic variations: their pattern and possible cause. Quaternary Research 3: 155–205. Dercourt, J., Ricou, L. E., and Vrielanck, B. (eds.) (1993), Atlas Tethys Paleoenvironmental Maps. Gauthier-Villars, Paris. Digerfeldt, G., Olson, S., and Sandgren, P. (2000), Reconstruction of lake-level changes in lake Xinias, central Greece, during the last 40,000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 65–82. Drivaliari, A. (1993), Image polliniques et paléoenvironments au Néogène supérieur en Méditerranée orientale. Aspects climatiques et paléo-géographiques d’un transect latitudinal (de la Roumanie au delta du Nil). Ph.D. Thesis, University of Montpellier. Dynesius, M. and Jansson, R. (2000), Evolutionary consequences of changes in species’ geographical distributions driven by Milankovitch climate oscillations. Proceedings of the National Academy of Sciences 97: 9115–20. Ehleringer, J. R. and Björkman, O. (1977), Quantum yields for CO2 uptake in C3 and C4 plants. Plant Physiology, 59: 86–90. Sage, R. F., Flanagan, L. B., and Pearcy, R. W. (1991), Climate change and the evolution of C4 photosynthesis. Trends in Ecology and Evolution 6: 95–9.

Cenozoic Climate and Vegetation Change Ehlers, J. (1996), Quaternary and Glacial Geology. Wiley, Chichester. Elenga, H., Peyron, O., Bonnefille, R., Jolly, D., Cheddadi, R., Guiot, J., Andrieu, V., Bottema, S., Buchet, G., de Beaulieu, J.L., Hamilton, A. C., Maley, J., Marchant, R., Perez-Obiol, R., Reille, M., G. Riollet, G., Scott, L., Straka, H., Taylor, D., and Van Campo, E. (2000), Pollen-based biome reconstruction for southern Europe and Africa 18,000 yr BP. Journal of Biogeography 27: 621–34. Elhaï, H. (1969), La flore sporo-pollinique du gisement Villafranchien de Senèze (Massif Central, France). Pollen et Spores 11: 127–39. Fairbanks, R. G. (1989), A 17,000 year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 442: 637–42. Mortlock, R. A., Chiu, T.-C., Cao, L., Kaplan, A., Guilderson, T. P., Fairbanks, T. W., Bloom, A. L., Grootes, P. M., and Nadeau, M.-J. (2005), Radiocarbon calibration curve spanning 0–50,000 years BP based on paired 230 Th/234 U/238 U and 14 C dates on pristine corals. Quaternary Science Reviews 24: 1781–96. Fjeldså, J. and Lovett, J.C. (1997), Geographical patterns of old and young species in African forest biota: the significance of specific montane areas as evolutionary centres. Biodiversity and Conservation 6: 325–46. Flohn, H. (1978), Comparison of Antarctic and Arctic climate and its relevance to climatic evolution, in E. M. van Zinderen Bakker (ed.), Antarctic Glacial History and World Palaeoenvironments. A. A. Balkema, Rotterdam, 3–13. Florschütz, F., Menéndez Amor, J., and Wijmstra, T. A. (1971), Palynology of a thick Quaternary succession in southern Spain. Palaeogeography, Palaeoclimatology, Palaeoecology 10: 233–64. Follieri, M., Magri, D., and Sadori, L. (1988), 250,000-year pollen record from Valle di Castiglione (Roma). Pollen et Spores 30: 329–56. Follieri, M., Giardini, M., Magri, D., and Sadori, L. (1998), Palynostratigraphy of the last glacial period in the volcanic region of central Italy. Quaternary International 47/8: 3–20. Francis, J. E. (1984), The seasonal environment of the Purbeck (Upper Jurassic) fossil forest. Palaeogeography, Palaeoclimatology, Palaeoecology 48: 285–307. Frenzel, B. (1968), The Pleistocene vegetation of northern Eurasia. Science 161: 637–49. (1979), Europe without forest. Geographical Magazine August 1979: 756–61. Frogley, M. R., Tzedakis, P. C., and Heaton, T. H. E. (1999), Climate variability in NW Greece during the last interglacial. Science 285: 1886–9. Fronval, T. and Jansen, E. (1997), Eemian and early Weichselian(140–60 ka) paleoceanography and paleoclimate in the Nordic seas with comparisons to Holocene conditions. Paleoceanography 12: 443–62. Galanidou, P. and Tzedakis, P. C. (2001), New AMS dates from the Upper Palaeolithic Kastritsa. Proceedings of the Prehistoric Society 67: 271–8. Lawson, I. T., and Frogley, M. R. (2000), A revised chronological and palaeoenvironmental framework for the Kastritsa rockshelter, northwest Greece. Antiquity 74: 349–55. Ganopolski, A. and Rahmstorf, S. (2001), Rapid changes of glacial climate simulated in a coupled climate model. Nature 409: 153–8.

131

Grivet, D. and Petit, R. J. (2003), Chloroplast DNA phylogeography of the hornbeam in Europe: Evidence for a bottleneck at the outset of postglacial colonization. Conservation Genetics 4: 47– 56. Haffer, J. (1969), Speciation in Amazonian forest birds. Science 165: 131–7. Haigh, J. D. (1996), The impact of solar variability on climate. Science 272: 981–4. Hampe, A. and Petit, R. J. (2005), Conserving biodiversity under climate change: the rear edge matters. Ecology Letters 8: 461–7. Haq, B. U., Hardenbol, J., and Vail, P. R. (1987), Chronology of fluctuating sea levels since the Triassic. Science 235: 1156–67. Harrison, S. P. and Digerfeldt, G. (1993), European lakes as palaeohydrological and palaeoclimatic indicators. Quaternary Science Reviews 12: 233–48. Yu, G., and Tarasov, P. E. (1996), Late Quaternary lakelevel record from northern Eurasia. Quaternary Research 45: 138–59. Haug, G. H. and Tiedemann, R. (1998), Effect of the formation of the Isthmus of Panama on Atlantic ocean thermohaline circulation. Nature 401: 777–82. Zahn, R., and Ravelo, A. C. (2001), Role of Panama uplift on oceanic freshwater balance. Geology 29: 207–10. Ganopolski, A., Sigman, D. M., Rosell-Mele, A., Swann, G. E. A., Tiedemann, R., Jaccard, S. L., Bollmann, J., Maslin, M. A., Leng, M. J., and Eglinton, G. (2005), North Pacific seasonality and the glaciation of North America 2.7 million years ago. Nature 433: 821–5. Hayes, A., Kucera, M., Kallel, N., Sbaffi, L., and Rohling, E. (2005), Glacial Mediterranean sea surface temperatures based on planktonic foraminiferal assemblages. Quaternary Science Reviews 24: 999–1016. Hays, J. D., Imbrie, J., and Shackleton, N. J. (1976), Variations in the Earth’s orbit: pacemaker of the ice ages. Science 194: 1121–31. Hazan, N., Stein, M., Agnon, A., Marco, S., Nadel, D., Negendank, J. F. W., Schab, M. J., and Neev, D. (2005), The late Quaternary limnological history of Lake Kinneret (Sea of Galille), Israel. Quaternary Research 63: 60–77. Hewitt, G. M. (1996), Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58: 247–76. Hewitt, G. M. (1999), Post-glacial re-colonization of European biota. Biological Journal of the Linnean Society 68: 87–112. (2000), The genetic legacy of the Quaternary ice ages. Nature 405: 907–13. Higgs, E. S., Vita-Finzi, C., Harris, D. R., and Fagg, A. E. (1967), The climate, environment and industries of Stone Age Greece, Part III. Proceedings of the Prehistoric Society 33: 1–29. Hilgen, F. J. (1991), Astronomical calibration of the Gauss to Matuyama sapropels in the Mediterranean and implications for the geomagnetic polarity time scale. Earth and Planetary Science Letters 104: 226–44. Lourens, L. J., Berger, A., and Loutre, M. F. (1993), Evaluation of the astronomically calibrated time scale for the late Pliocene and earliest Pleistocene. Paleoceanography 8: 549–65. Holzkämper, S., Mangini, A., Spötl, C., and Mudelsee, M. (2004), Timing and progression of the Last Interglacial derived from a high alpine stalagmite. Geophysical Research Letters 31: L07201, doi: 1029/2003GL019112. Horowitz, A. (1974), Some pollen spectra from the Neogene of Israel. Pollen et Spores 16: 59–65.

132

Chronis Tzedakis

Horrell, M. A. (1991), Phytogeography and paleoclimatic interpretation of the Maestrichtian. Palaeogeography, Palaeoclimatology, Palaeoecology 86: 87–138. Hsü, K. J., Montadert, L., Bernoulli, D., Cita, M. B., Erickson, A., Garrison, R. E., Kidd, R. B., Mèlierés, F., Müller, C., and Wright, R. (1977), History of the Mediterranean salinity crisis. Nature 267: 399–403. Hughes, P. D., Gibbard, P. L., and Woodward, J. C. (2003), Relict rock glaciers as indicators of Mediterranean palaeoclimate during the Last Glacial Maximum (Late Würmian) in northwest Greece. Journal of Quaternary Science 18: 431–40. Woodward, J. C., Gibbard, P. L., Macklin, M. G., Gilmour, M. A., and Smith, G. R. (2006), Quaternary glacial history of the Pindus Mountains, Greece. Journal of Geology 114: 413–34. Huntley, B. (1993), Species richness in north-temperate zone forests. Journal of Biogeography 20: 163–80. and Birks, H. J. B. (1983), An Atlas of Past and Present Pollen Maps for Europe: 0–13000 Years Ago. Cambridge University Press, Cambridge. and Webb, T., III (1989), Migration: species’ response to climatic variations caused by changes in the Earth’s orbit. Journal of Biogeography 16: 5–19. Imbrie, J., Hays, J. D., Martinson, D. G., McIntyre, A., Mix, A. C., Morley, J. J., Pisias, N. G., Prell, W. L., and Shackleton, N. J. (1984), The orbital theory of Pleistocene climate: support from a revised chronology of the marine ‰O18 record, in A. L. Berger, J. Imbrie, J. Hays, G. Kukla, and B. Saltzman (eds.), Milankovitch and Climate. D. Reidel, Dordrecht, 269–306. Iversen, J. (1958), The bearing of glacial and interglacial epochs on the formation and extinction of plant taxa. Uppsala Universitets Årsskrift 6: 210–15. Jacobs, B. F. (2004), Palaeobotanical studies from tropical Africa: relevance to the evolution of forest, woodland and savannah biomes. Philosophical Transactions of the Royal Society London B 359: 1573–83. Jacobs, E., Weissert, G., Shields, G., and Stille, P. (1996), The Monterey event in the Mediterranean: A record from shelf sediments of Malta. Paleoceanography 11: 717–28. Janis, C. M. (1993), Tertiary mammal evolution in the context of changing climates, vegetation and tectonic events. Annual Review of Ecology and Systematics 24: 467–500. Jolly, D., Harrison, S. P., Damnati, B., and Bonnefille, R. (1998), Simulated climate and biomes of Africa during the Late Quaternary: comparison with pollen and lake status data. Quaternary Science Reviews 17: 629–57. Kallel, N., Paterne, M., Duplessy, J. C., Vergnaud-Grazzini, C., Pujol, C., Labeyrie, L., Arnold, M., Fontugne, M., and Pierre, C. (1997), Enhanced rainfall in the Mediterranean region during the last sapropel event. Oceanologica Acta 20: 697–712. Keeley, J. E. and Rundel, P. W. (2005), Fire and the Miocene expansion of C4 grasslands. Ecology Letters, 8: 683–90. Kellogg, E. A. (2001), Evolutionary history of the grasses. Plant Physiology 125: 1198–205. Kennett, J. P. (1995), A review of polar climatic evolution during the Neogene based on the marine sediment record, in E. S. Vrba, G. H. Denton, T. C. Partridge, and L. H. Burckle (eds.), Paleoclimate and Evolution. Yale University Press, New Haven, 49–64. King, G. C. P. and Bailey, G. N. (1985), The palaeoenvironment of some archaeological sites in Greece: the influence of accumulated uplift in a seismically active region. Proceedings of the Prehistoric Society 51: 273–83.

King, R. A. and Ferris, C. (1998), Chloroplast DNA phylogeography of Alnus glutinosa (L.) Gaertn. Molecular Ecology 7: 1151–61. Kingston, J. D., Marino, B. D., and Hill, A. (1994), Isotopic evidence for Neogene hominid paleoenvironments in the Kenya Rift Valley. Science 264: 955–9. Klicka, J. and Zink, R. M. (1997), The importance of recent ice ages in speciation: a failed paradigm. Science 277: 1666–9. Kloosterboer-van Hoeve, M. (2000), Cyclic Changes in the Late Neogene Vegetation of Northern Greece. Ph.D. Thesis, University of Utrecht. Klotz, S., Fauquette, S., Combourieu-Nebout, N., Uhl, D., Suc, J.-P., and Mosbrugger, V. (2006), Seasonality intensification and long-term winter cooling as part of the Late Pliocene climate development. Earth and Planetary Science Letters 241: 174–87. Krijgsman, W. (2002), The Mediterranean: Mare Nostrum of Earth sciences. Earth and Planetary Science Letters 205: 1–12. Hilgen, F. J., Raffi, I., Sierro, F. J., and Wilson, D. S. (1999), Chronology, causes and progression of the Messinian salinity crisis. Nature 400: 652–5. Krom, M. D., Stanley, J. D., Cliff, R. A., and Wooward, J. C. (2002), Nile River sediment fluctuations over the past 7000 yr and their key role in sapropel development. Geology 30: 71–4. Kuper, R. and Kröpelin, S. (2006), Climate-controlled Holocene occupation in the Sahara: Motor of Africa’s evolution. Science 313: 803–7, doi: 10.1126/science.1130989. Lagercrantz, U. and Ryman, N. (1990), Genetic structure of the Norway spruce (Picea abies): concordance of morphological and allozymic variation. Evolution 44: 38–53. Lamb, H. F., Eicher, U., and Switsur, V. R. (1989), An 18,000year record of vegetation, lake-level and climatic change from Tigalmamine, Middle Atlas, Morocco. Journal of Biogeography 16: 65–74. Gasse, F., Benkaddour, A., El Hamouti, N., van der Kaars, S., Perkins, W. T., Pearce, N. J., and Roberts, C. N. (1995), Relation between century-scale Holocene arid intervals in tropical and temperate zones. Nature 373: 134–7. Lang, G. (1970), Florengeschichte und mediterranmitteleuropäische Florenbeziehungen. Feddes Repertorium, 81: 315–55. Larcher, W. (1981), Low temperature effects on Mediterranean sclerophylls: an unconventional viewpoint, in N. S. Margaris and H. A. Mooney (eds.), Components of Productivity of Mediterranean-Climate Regions—Basic and Applied Aspects. Dr W. Junk, The Hague, 259–66. Larrasoaña, J. C., Roberts, A. P., Rohling, E. J., Winklhofer, M., and Wehausen, R. (2003), Three million years of monsoon variability over the northern Sahara. Climate Dynamics 21: 689–98. Lascoux, M., Palmé, A. E., Cheddadi, R., and Latta, R. G. (2004), Impact of Ice Ages on the genetic structure of trees and shrubs. Philosophical Transactions of the Royal Society London B 359: 197–207. Lawson, I. T., Frogley, M. R., Tzedakis, P. C., Preece, R. C., and Bryant, C. (2004), Lateglacial and Holocene environmental history of the Ioannina basin, north-west Greece. Quaternary Science Reviews, 23: 1599–625. Lehman, S., Sachs, J. P., Crotwell, A. M., Keigwin, L. D., and Boyle, E. A. (2002), Relation of subtropical Atlantic temperature, high-latitude ice rafting, deep water formation, and European climate 130,000–60,000 years ago. Quaternary Science Reviews 21: 1917–24.

Cenozoic Climate and Vegetation Change Leroy, S. A. G. and Seret, G. (1996), Duration and vegetation dynamics of the Nogaret interglacial (∼1.9 Ma) (S. of France)—tentative correlation with Stage 75, in G. Kukla and E. Went (eds.), Start of a Glacial. Springer, Heidelberg, 113–25. Livermore, R. A. and Smith, A. G. (1985), Some boundary conditions for the evolution of the Mediterranean region in D. J. Stanley and F. C. Wezel (eds.), Geological Evolution of the Mediterranean Basin. Springer, New York, 83–98. Lourens, L. J., Hilgen, F. J., Gudjonsson, L., and Zachariasse, W. J. (1992), Late Pliocene to early Pleistocene astronomically forced sea surface productivity and temperature variations in the Mediterranean. Marine Micropaleontology 19: 49–78. Antonarakou, A., Hilgen, F. J., Van Hoof, A. A. M., Vergnaud-Grazzini, C., and Zachariasse, W. J. (1996), Evaluation of the Plio-Pleistocene astronomical timescale. Paleoceanography 11: 391–413. and Hilgen, F. J. (1997), Long-periodic variations in the Earth’s obliquity and their relation to third-order eustatic cycles and Late Neogene glaciations. Quaternary International, 40: 43–52. Wehausen, R., and Brumsack, H. J. (2001), Geological constraints on tidal dissipation and dynamical ellipticity of the Earth over the past three million years. Nature 409: 1029–33. Hilgen, F., Shackleton, N. J., Laskar, J., and Wilson, D. (2004), The Neogene Period, in F. M. Gradstein, J. G. Ogg, and A. G. Smith, (eds.), A Geologic Time Scale. Cambridge University Press, Cambridge, 409–41. Lunt, D. J., Valdes, P. J., Haywood, A., and Rutt, I. C. (2008), Closure of the Panama Seaway during the Pliocene: implications for climate and Northern Hemisphere glaciation. Climate Dynamics 30: 1–18. McKenzie, J. (1999), From desert to deluge in the Mediterranean. Nature 400: 613–14. Spezzaferri, S., and Isern, A. (1999), ‘The Miocene/Pliocene Boundary in the Mediterranean Sea and Bahamas.: Implications for a global flooding event in the Earliest Pliocene. Memoire Società Geologica Italiana 54: 93–108. McManus, J. F., Oppo, D. W., and Cullen, J. L. (1999), A 0.5million-year record of millennial-scale climate variability in the North Atlantic. Science 283: 971–5. François, R., Gherardi, J.-M., Keigwin, L. D., and BrownLeger, S. (2004), Collapse and rapid resumption of Atlantic meridional circulation linked to deglacial changes. Nature 428: 834–7. Magri, D. (1999), Late Quaternary vegetation history at Lagaccione near Lago di Bolsena (central Italy). Review of Palaeobotany and Palynology 106: 171–208. and Parra, I. (2002), Late Quaternary western Mediterranean pollen records and African winds. Earth and Planetary Science Letters 200: 401–8. and Sadori, L. (1999), Late Pleistocene and Holocene pollen stratigraphy at lago di Vico, central Italy. Vegetation History and Archaeobotany 8: 247–60. and Tzedakis, P. C. (2000), Orbital signatures and long-term vegetation patterns in the Mediterranean. Quaternary International 73/4: 69–78. Kallel, N., and Narcisi, B. (2004), Palaeoenvironmental changes in the Mediterranean region 250–10 kyr BP, in R. W. Battarbee et al. (eds.), Past Climate Variability through Europe and Africa. Kluwer, Dordrecht, 325–41. Mai, D. H. (1989), Development and regional differentiation of the European vegetation during the Tertiary. Plant Systematics and Evolution 162: 79–91.

133

Mandell, R. D. and Simmons, A. H. (2001), Prehistoric occupation of Late Quaternary landscapes near Kharga Oasis, western desert of Egypt. Geoarchaeology 16: 95–117. Martinson, D. G., Pisias, N. G., Hays, J. D., Imbrie, J., Moore Jr., T. C., and Shackleton, N. J. (1987), Age dating and the orbital theory of ice ages: development of a high-resolution 0 to 300,000-year chronostratigraphy. Quaternary Research 27: 1–29. Mayewski, P. A., Meeker, L. D., Twickler, M. S., Whitlow, S., Yang, Q., Lyons, W. B., and Prentice, M. (1997), Major features and forcing of high-latitude northern hemisphere atmospheric circulation using a 110,000-year-long glaciochemical series. Journal of Geophysical Research 102 (C12): 26,345–66. Meadows, J. (2005), The Younger Dryas episode and the radiocarbon chronologies of the Lake Huleh and Ghab Valley pollen diagrams, Israel and Syria. The Holocene 15: 631–6. Middleton, N. J. (1985), Effect of drought on dust production in the Sahel. Nature 316: 431–4. Milankovitch, M. (1930), Mathematische Klimalehre und astronomische Theorie der Klimaschwankungen. In W. Köppen, and R. Geiger (eds.), Handbuch der Klimatologie, I (A). Gebrüder Borntraeger, Berlin, 1–176. Mix, A., Bard, E., and Schneider, R. (2001), Environmental processes of the ice age: land, oceans, glaciers (EPILOG). Quaternary Science Reviews 20: 627–57. Moreno, A., Cacho, I., Canals, M., Grimalt, J. O., Sánchez Goñi, M. F., Shackleton, N., and Sierro, F. J. (2005), Links between marine and atmospheric processes oscillating on a millennial time-scale. A multi-proxy study of the last 50,000 yr from the Alboran Sea (Western Mediterranean). Quaternary Science Reviews 24: 1623–36. Mudelsee, M. and Schulz, M. (1997), The Mid-Pleistocene climate transition: onset of 100 ka cycle lags ice volume build-up by 280 ka. Earth and Planetary Science Letters 151: 117–23. Muñoz Sobrino, C., Ramil-Rego, P., Cómez-Ortellana, L., and Díaz Varela, R. A. (2005), Palynological data on major Holocene climatic events in NW Iberia. Boreas 34: 381–400. Muttoni, G., Ravazzi, C., Breda, M., Pini, R., Laj, C., Kissel, C., Mazaud, A., and Garzanti, E. (2007), Magnetostratigraphic dating of an intensification of glacial activity in the southern Italian Alps during Marine Isotope Stage 22. Quaternary Research 67: 161–73. Nambudiri, E. M. V., Tidwell, W. D., Smith, B. N., and Hebbert, N. P. (1978), A C4 plant from the Pliocene. Nature 276: 816–17. Neev, D. and Emery, K. O. (1967), The Dead Sea; depositional processes and environments of evaporites. Geological Survey of Israel Bulletin 41. Niklewski, J. and van Zeist, W. (1970), A Late Quaternary pollen diagram from northwestern Syria. Acta Botanica Neerlandica, 19: 737–54. Okuda, M., Yasuda, Y., and Setoguchi, T. (2001), Middle to Late Pleistocene vegetation history and climate changes at Lake Kopais, Southeast Greece. Boreas 30: 73–82. Pagani, M., Arthur, M. A., and Freeman, K. H. (1999a ), Miocene evolution of atmospheric carbon dioxide. Paleoceanography 14: 273–92. Freeman, K. H., and Arthur, M. A. (1999b), Late Miocene atmospheric CO2 concentrations and the expansion of C4 grasses. Science 285: 876–9. Zachos, J. C., Freeman, K. H., Tipple, B., and Bohaty, S. (2005), Marked decline in atmospheric carbon dioxide concentrations during the Paleogene. Science 309: 600–3.

134

Chronis Tzedakis

Palamarev, E. (1989), Paleobotanical evidences of the Tertiary history and origin of the Mediterranean sclerophyll dendroflora. Plant Systematics and Evolution 162: 93–107. Palmé, A. E. and Vendramin, G. G. (2002), Chloroplast DNA variation, postglacial recolonization and hybridization in hazel, Corylus avellana. Molecular Ecology 11: 1769–79. Semerikov, V., and Lascoux, M. (2003a ), Absence of geographical structure of chloroplast DNA in sallow Salix capreas. Heredity 91: 465–74. Su, Q., Rautenberg, A., Manni, F., and Lascoux, M. (2003b), Postglacial recolonization and cpDNA variation of silver birch, Betula pendula. Molecular Ecology 12: 201–12. Pantaléon-Cano, J., Yll, E.-I., Perez-Obiol, R., and Roure, J. M. (2003), Palynological evidence for vegetational history in semi-arid areas of the western Mediterranean (Almería, Spain). The Holocene 13: 109–19. Pérez-Obiol, R. and Julià, R. (1994), Climatic change on the Iberian Peninsula recorded in a 30,000-yr pollen record from Lake Banyoles. Quaternary Research 41: 91–8. Petit, R. J., Brewer, S., Bordacs, S., Burg, K., Cheddadi, R., Coart, E., Cottrell, J., Csaikl, U. M., van Dam, B., Deans, J. D., Espinel, S., Fineschi, S., Finkeldey, R., Glaz, I., Goicoechea, P. G., Jensen, J. S., Konig, A. O., Lowe, A. J., Madsen, S. F., Matyas, G., Munro, R. C., Popescu, F., Slade, D., Tabbener, H., de Vries, S. G. M., Ziegenhagen, B., de Beaulieu, J.-L., and Kremer, A. (2002), Identification of refugia and post-glacial colonisation routes of European white oaks based on chloroplast DNA and fossil evidence. Forest Ecology and Management 156: 49–74. Aguinagalde, I., de Beaulieu, J.-L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, R., Fineschi, S., Grivet, D., Lascoux, M., Mohanty, A., Muller-Starck, G. M., Demesure-Musch, B., Palme, A., Martin, J. P., Rendell, S., Vendramin, G. G. (2003), Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 1563–5. Pollard, D. and Barron, E. J. (2003), Causes of model-data discrepancies in European climate during oxygen isotope stage 3 with insights from the Last Glacial Maximum. Quaternary Research 59: 108–13. Pons, A. and Reille, M. (1988), The Holocene- and Upper Pleistocene pollen record from Padul (Granada, Spain): a new study. Palaeogeography, Palaeoclimatology, Palaeoecology 66: 24–63. Prasad, S., Vos, H., Negendank, J. F. W., Waldmann, N., Goldstein, S. L., and Stein, M. (2004), Evidence from Lake Lisan of solar influence on decadal- to centennial-scale climate variability during marine oxygen isotope stage 2. Geology 32: 581–4. Prentice, I. C., Guiot, J., and Harrison, S. P. (1992), Mediterranean vegetation, lake levels and palaeoclimate at the Last Glacial Maximum. Nature 360: 658–60. Jolly, D., and BIOME 6000 participants (2000), MidHolocene and glacial-maximum vegetation geography of the northern continents and Africa. Journal of Biogeography 27: 507–19. Quade, J., Solounias, N., and Cerling, T. E. (1994), Stable isotopic evidence from paleosol carbonates and fossil teeth in Greece for forest or woodlands over the past 11 Ma. Palaeogeography, Palaeoclimatology, Palaeoecology 108: 41–53. Rahmstorf, S. (2002), Ocean circulation and climate during the past 120,000 years. Nature 419: 207–14. (2003), Timing of abrupt climate change: a precise – clock. Geophysical Research Letters 30/10: 1510, doi: 10.1029/2003GL017115.

Ravazzi, C. and Rossignol-Strick, M. (1995), Vegetation change in a climatic cycle of Early Pleistocene age in the Leffe Basin (Northern Italy). Palaeogeography, Palaeoclimatology, Palaeoecology 117: 105–22. Raven, P. H. (1973), The evolution of Mediterranean Floras, In F. di Castri and H. A. Mooney (eds.), Mediterranean Type Ecosystems. Chapman & Hall, London, 213–24. Raymo, M. E., Ruddiman, W. F., Backman, J., Clement, B. M., and Martinson, D. G. (1989), Late Pliocene variation in northern hemisphere ice sheets and north Atlantic deep water circulation. Paleoceanography, 4: 413–46. Ganley, K., Carter, S., Oppo, D. W., and McManus, J. (1998), Millennial-scale climate instability during the early Pleistocene epoch. Nature 394: 699–702. Reed, J. M., Stevenson, A. C., and Juggins, S. (2001), A multiproxy record of Holocene climatic change in southwestern Spain: the Laguna de Medina, Cádiz. The Holocene 11: 707–19. Reid, E. M. (1935), British floras antecedent to the Great Ice Age. Discussion on the origin and relationship of the British Flora. Proceedings of the Royal Society of London B 118: 197. Reille, M., and de Beaulieu, J.-L. (1990), Pollen analysis of a long Upper Pleistocene continental sequence in a Velay maar (Massif Central, France). Palaeogeography, Palaeoclimatology, Palaeoecology 80: 35–48. (1995), Long Pleistocene pollen records from the Praclaux Crater, south-central France. Quaternary Research 44: 205–15. and Lowe, J. J. (1993), A re-evaluation of the vegetation history of the Eastern Pyrenees (France) from the end of the last glacial to the present. Quaternary Science Reviews 12: 47–77. Reille, M., Andrieu, V., de Beaulieu, J.-L., Guenet, P., and Goeury, C. (1998), A long pollen record from Lac du Bouchet, Massif Central, France for the period 325 to 100 ka (OIS 9c to OIS 5e). Quaternary Science Reviews 17: 1107–23. Ritchie, J. C. and Haynes, C. V. (1987), Holocene vegetation zonation in the eastern Sahara. Nature 330: 645–7. Roberts, N. and Wright Jr., H. E. (1993), Vegetational, lake level and climatic history of the Near East and Southwest Asia, in, H. E. Wright Jr., J. E. Kutzbach, T. Webb III, W. F. Ruddiman, F. A. Street-Perrot, and P. J. Bartlein (eds.), Global Climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis, 194–220. Roger, S., Coulon, C., Thouveny, N., Féraud, G., Van Velzen, A., Fauquette, S., Cochemé, J. J., Prévot, M., and Verosub, K. L. (2000), 40 Ar/39 Ar dating of a tephra layer in the Pliocene Senèze maar lacustrine sequence (French Massif Central): constraint on the age of the Réunion-Matuyama transition and implications on palaeoenvironmental archives. Earth and Planetary Science Letters 183: 431–40. Rohling, E. J. and Hilgen, F. J. (1991), The eastern Mediterranean climate at times of sapropel formation: a review. Geologie en Mijnbouw 70: 77–83. Cane, T. R., Cooke, S., Sprovieri, M., Bouloubassi, I., Emeis, K. C., Schiebel, R., Kroon, D., Jorissen, F. J., Lorre, A., and Kemp, A. E. S. (2002a ), African monsoon variability during the previous interglacial maximum. Earth and Planetary Science Letters 202: 61–75. Mayewski, P. A., Abu-Zied, R. H., Casford, J. S. L., and Hayes, A. (2002b), Holocene atmosphere-ocean interactions: Records from Greenland and the Aegean Sea. Climate Dynamics 18: 587–93.

Cenozoic Climate and Vegetation Change Mayewski, P. A., and Challenor, P. (2003), On the timing and mechanism of millennial-scale climate variability during the last glacial cycle. Climate Dynamics 20: 257–67. Rossignol-Strick, M. (1983), African monsoons, an immediate climate response to orbital insolation. Nature 304: 46–9. (1995), Sea-land correlation of pollen records in the Eastern Mediterranean for the glacial–interglacial transition: biostratigraphy versus radiometric dating. Quaternary Science Reviews 14: 893–915. (1999), The Holocene climatic optimum and pollen records of sapropel 1 in the eastern Mediterranean, 9000–6000 BP. Quaternary Science Reviews, 8: 515–30. and Paterne, M. (1999), A synthetic pollen record of the eastern Mediterranean sapropels of the last 1 Ma: implications for the time-scale and formation of sapropels. Marine Geology 153: 221–37. Nesteroff, W., Olive, P., and Vergnaud-Grazzini, C. (1982), After the deluge: Mediterranean stagnation and sapropel formation. Nature 295: 105–10. Roucoux, K. H., Shackleton, N. J., de Abreu, L., Schonfeld, J., and Tzedakis, P. C. (2001), Combined marine proxy and pollen analyses reveal rapid Iberian vegetation response to North Atlantic millennial-scale climate oscillations. Quaternary Research 56: 128–32. de Abreu, L., Shackleton, N. J., and Tzedakis, P. C. (2005), The response of NW Iberian vegetation to North Atlantic climate oscillations during the last 65 kyr. Quaternary Science Reviews 25: 1637–53. Roy, K., Valentine, J. W., Jablonski, D., and Kidwell, S. M. (1996), Scale of climatic variability and time averaging in Pleistocene biotas: implications for ecology and evolution. Trends in Ecology and Evolution 11: 458–63. Sadori, L. and Narcisi, B. (2001), The Postglacial record of environmental history from Lago de Pergusa, Sicily. The Holocene 11: 655–70. Salamani, M. (1993), Prémières données paléophytogéographiques du Cèdre de l’Atlas (Cedrus atlantica) dans la région de Grand Kabylie (NE Algérie). Palynosciences 2: 147–55. Sánchez Goñi, M. F., Turon, J.-L., Eynaud, F., and Gendreau, S. (2000), European climatic response to millennial-scale changes in the atmosphere-ocean system during the last glacial period. Quaternary Research 54: 394–403. Sierro, F. J., Peypouquet, J.-P., Grimalt, J. O., Shackleton, N. J., Cacho, I., Turon, J.-L., and Guiot, J. (2002), Synchroneity between marine and terrestrial responses to millennial scale climatic variability during the last glacial period in the Mediterranean region. Climate Dynamics 19: 95–105. Sbaffi, L., Wezel, F. C., Curzi, G., and Zoppi, U. (2004), Millennialto centennial-scale palaeoclimatic variations during Termination I and the Holocene in the central Mediterranean Sea. Global and Planetary Change 40: 201–17. Scrivner, A. E., Vance, D., and Rohling, E. J. (2004), New neodymium isotope data quantify Nile involvement in Mediterranean anoxic episodes. Geology 32: 565–8. Sercelj, A. (1966), Pelodne analize Pleistocenskih in Holocenskih sedimentov Ljubljanskega barga. Razprave IX, Slovenska Akademija Znanosti Utmentnosit 9: 431–72. Shackleton, N. J. (1969), The last interglacial in the marine and terrestrial records. Proceedings of the Royal Society, London B 174: 135–54. (1987), Oxygen isotopes, ice volume and sea level. Quaternary Science Reviews 6: 183–90.

135

(1995), New data on the evolution of Pliocene climatic variability, In E. S. Vrba, G. H. Denton, T. C. Partridge, and L. H. Burckle (eds.), Paleoclimate and Evolution. Yale University Press, New Haven, 242–8. (1997), The deep-sea sediment record and the PlioPleistocene boundary. Quaternary International 40: 33–5. (2001), Milankovitch in the Miocene. EOS Transactions American Geophysical Union, 82(47) Fall Meeting Supplement, Abstract U11A-12. and Kennett, J. P. (1975), Paleotemperature history of the Cenozoic and the initiation of Antarctic glaciation: Oxygen and carbon isotope analyses in DSDP Sites 277, 279 and 281, in J. P. Kennett and R. E. Houtz, (eds.), Initial Reports of the Deep Sea Drilling Project 29: 743–56. Imbrie, J., and Hall, M. A. (1983), Oxygen and carbon isotope record of East Pacific core V19-30: implications for the formation of deep water in the late Pleistocene North Atlantic. Earth and Planetary Science Letters 65: 233–44. Backman, J., Zimmerman, H., Kent, D. V., Hall, M. A., Roberts, D. G., and Baldauf, J. (1984), Oxygen isotope calibration of the onset of ice-rafting and history of glaciation in the North Atlantic region. Nature 307: 620–3. Berger, A., and Peltier, W. R. (1990), An alternative astronomical calibration of the Lower Pleistocene timescale based on ODP Site 677. Transactions of the Royal Society of Edinburgh: Earth Sciences 81: 251–61. Hall, M. A., and Pate, D. (1995a ), Pliocene stable isotopic stratigraphy of Site 846. In N. G. Pisias, L. A. Mayer, T. R. Janecek, and T. H. van Andel (eds.), Proceedings of the Ocean Drilling Program, Scientific Results 138: 337–55. Crowhurst, S., Hagelberg, T., Pisias, N. G., and Schneider, D. A. (1995b), A new Late Neogene time scale: application to Leg 138 Sites, in N. G. Pisias, L. A. Mayer, L. A. Janacek, and T. H. van Andel (eds.), Proceedings of the Ocean Drilling Program, Scientific Results, 138: 73–101. Hall, M. A. and Vincent, E. (2000), Phase relationships between millennial scale events 64,000–24,000 years ago. Paleoceanography 15: 565–9. Chapman, M., Sánchez-Goñi, M. F., Pailler, D., and Lancelot, Y. (2002), The classic Marine Isotope Substage 5e. Quaternary Research 58: 14–16. (2003), Marine Isotope Substage 5e and the Eemian Interglacial. Global and Planetary Change 36: 151–5. Shinn, R. A. and Barron, E. J. (1989), Climatic sensitivity to continental ice-sheet size and configuration. Journal of Climate 2: 1517–37. Sierro, F. J., Hodell, D. A., Curtis, J. H., Flores, J. A., Reguera, I., Colmenero-Hidalgo, E., Bárcena, M. A., Grimalt, J. O., Cacho, I., Frigola, J., and Canals, M. (2005), Impact of iceberg melting on the Mediterranean thermohaline circulation during Heinrich events. Paleoceanography, 20: PA2019, doi: 10.1029/2004PA001051. Singh, G. (1988), History of aridland vegetation and climate: a global perspective. Biological Reviews 63: 159–95. Smit, A. and Wijmstra, T. A. (1970), Application of transmission electron microscope analysis to the reconstruction of former vegetation. Acta Botanica Neerlandica, 19: 867–76. Smith, A. G., Smith, D. G., and Funnell, B. M. (1992), Atlas of Mesozoic and Cenozoic Coastlines. Cambridge University Press, Cambridge. Sprovieri, M., Sacchi, M., and Rohling, E. J. (2003), Climatically influenced interactions between the Mediterranean and the

136

Chronis Tzedakis

Paratethys during the Tortonian. Paleoceanography 18: 1034, doi: 10.1029/2001PA000750. Steenbrink, J., van Vugt, N., Hilgen, F. J., Wijbrans, J. R., and Meulenkamp, J. E. (1999), Sedimentary cycles and volcanic ash beds in the Lower Pliocene lacustrine succession of Ptolemais (NW Greece); discrepancy between 40 Ar/39 Ar and astronomical ages. Palaeogeography, Palaeoclimatology, Palaeoecology 152: 283–303. Stewart, J. R. (2003), Comment on ‘Buffered vegetation changes in a Quaternary refugium: evolutionary implications’. Science 299: 825a. and Lister, A. M. (2001), Cryptic northern refugia and the origins of the modern biota. Trends in Ecology and Evolution, 16: 608–13. Stuiver, M. and Braziunas, T. F. (1989), Atmospheric 14 C and century-scale solar oscillations. Nature, 338: 405–8. and Grootes, P. M. (2000), GISP2 Oxygen Isotope Ratios. Quaternary Research, 53: 277–84. Suc, J.-P. (1984), Origin and evolution of the Mediterranean vegetation and climate in Europe. Nature 307: 429–32. Diniz, F., Leroy, S., Poumot, C., Bertini, A., Dupont, L., Clet, M., Bessais, E., Zheng, Z., Fauquette, S., and Ferrier, J. (1995a ), Zanclean (∼Brunssumian) to early Piacenzian (∼ early-middle Reuverian) climate from 4◦ to 54◦ North latitude (West Africa, West Europe and West Mediterranean areas). Mededelingen Rijks Geologische Dienst 52: 43–56. Bertini, A., Combourieu-Nebout, N., Diniz, F., Leroy, S., Russo-Ermolli, F., Zheng, Z., Bessais, E., and Ferrier, J. (1995b), Structure of West Mediterranean vegetation and climate since 5.3 Ma. Acta Zoologica Cracoviense 38: 3–16. and Popescu, S.-M. (2005), Pollen records and climatic cycles in the Mediterranean region since 2.7 Ma, in M. J. Head and P. L. Gibbard (eds.), Early-Middle Pleistocene Transitions: The Land–Ocean Evidence. Geological Society of London Special Publication, London 247, 147–58. Svenning, J.-C. (2003), Deterministic Plio-Pleistocene extinctions in the European cool-temperate tree flora. Ecology Letters 6: 646–53. Taberlet, P., Fumagalli, L., Wust-Saucy, A.-G., and Cosson, J.F. (1998), Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 7: 453–64. Tiedemann, R., Sarnthein, M., and Shackleton, N. J. (1994), Astronomic timescale for the Pliocene Atlantic ‰18 O and dust flux records of Ocean Drilling Program site 659. Paleoceanography 9: 619–38. Tinner, W. and Lotter, A. F. (2001), Central European vegetation response to abrupt climate change at 8.2 ka. Geology 29: 551–4. Triat-Laval, H. (1978), Contribution pollenanalytique à l’histoire tardiglaciaire et holocène de la végétation de la basse vallée du Rhône. Thesis, Aix-Marseille 3 University. Tuenter, E., Weber, S. L., and Hilgen, L. J. (2003), The response of the African summer monsoon to remote and local forcing due to precession and obliquity. Global and Planetary Change 772: 1–17. Turco, E., Hilgen, F. J., Lourens, L. J., Shackleton, N. J., and Zachariasse, W. J. (2001), Punctuated evolution of global climate cooling during the late Middle to early Miocene: High-resolution planktonic foraminiferal and oxygen isotope records from the Mediterranean. Paleoceanography 16: 405–23. Turner, C. and West, R. G. (1968), The subdivision and zonation of interglacial periods. Eiszeitalter und Gegenwart 19: 93–101.

Tzedakis, P. C. (1993), Long-term tree populations in northwest Greece through multiple Quaternary climatic cycles. Nature 364: 437–40. (1994), Vegetation change through glacial-interglacial cycles: a long pollen sequence perspective. Philosophical Transactions of the Royal Society of London B, 345: 403–32. (1999), The last climatic cycle at Kopais, central Greece. Journal of the Geological Society, London, 155: 425–34. (2005), Towards an understanding of the response of southern European vegetation to orbital and suborbital climate variability. Quaternary Science Reviews 25: 1585–99. and Bennett, K. D. (1995), Interglacial vegetation succession: a view from southern Europe. Quaternary Science Reviews 14: 967–82. Andrieu, V., de Beaulieu, J.-L., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N. J., and Wijmstra, T. A. (1997), Comparison of terrestrial and marine records of changing climate of the last 500,000 years. Earth and Planetary Science Letters 150: 171–6. Andrieu, V., Birks, H. J. B., de Beaulieu, J.-L., Crowhurst, S., Follieri, M., Hooghiemstra, H., Magri, D., Reille, M., Sadori, L., Shackleton, N. J., and Wijmstra, T. A. (2001), Establishing a terrestrial chronological framework as a basis for biostratigraphical comparisons. Quaternary Science Reviews 20: 1583–92. Frogley, M. R., and Heaton, T. H. E. (2002a ), Duration of last interglacial conditions in northwest Greece. Quaternary Research 58: 53–5. Lawson, I. T., Frogley, M. R., Hewitt, G. M., and Preece, R. C. (2002b), Buffered vegetation changes in a Quaternary refugium: evolutionary implications. Science 297: 2044–7. (2003a ), Reply to comment on ‘Buffered vegetation changes in a Quaternary refugium: evolutionary implications’. Science 299: 825b. Tzedakis, P. C., Frogley, M. R., and Heaton, T. H. E. (2003b), Last interglacial conditions in southern Europe: evidence from Ioannina, northwest Greece. Global and Planetary Change 36: 157–70. McManus, P. C., Hooghiemstra, H., Oppo, D. W., and Wijmstra, T. A. (2003c), Comparison of changes in vegetation in northeast Greece with records of climate variability on orbital and suborbital frequencies over the last 450,000 years. Earth and Planetary Science Letters 21: 197–212. Frogley, M. R., Lawson, I. T., Preece, R. C., Cacho, I., and de Abreu, L. (2004a ), Ecological thresholds and patterns of millennial-scale climate variability: The response of vegetation in Greece during the last glacial period. Geology 32: 109–12. Roucoux, K. H., de Abreu, L., and Shackleton, N. J. (2004b), The duration of forest stages in southern Europe and interglacial climate variability. Science 306: 2231–5, doi: 10.1126/ science. 1102398. Hooghiemstra, H., and Pälike, H. (2006), The last 1.35 million years at Tenaghi Philippon: revised chronostratigraphy and long-term vegetation trends. Quaternary Science Reviews 25: 3416–30. Vaks, A., Bar-Matthews, M., Ayalon, A., Schilman, B., Gilmour, M., Hawkesworth, C. J., Frumkin, A., Kaufman, A., and Matthews, A. (2003), Paleoclimate reconstruction based on the timing of speleothem growth and oxygen and carbon isotope composition in a cave located in the rain shadow in Israel. Quaternary Research, 59: 182–93.

Cenozoic Climate and Vegetation Change van Campo, M., Aymonin, G., Guinet, P., and Rognon, P. (1964), Contribution à l’étude de peuplement végétal Quaternaire des montagnes sahariennes: L’Atakor. Pollen et Spores 6: 169–94. Guinet, P., and Cohen, J. (1968), Fossil pollen from late Tertiary and middle Pleistocene deposits of the Kurkur Oasis, in K. W. Butzer and C. H. Hanssen (eds.), Desert and River in Nubia: Geomorphology and Prehistoric Environments at the Answan Reservoir, Appendix I. University of Wisconsin Press, Madison, 515–20. van der Hammen, T., Wijmstra, T. A., and Zagwijn, H. (1971), The floral record of the Late Cenozoic of Europe, in K. K. Turekian (ed.), The Late Cenozoic Glacial Ages. Yale University Press, New Haven, 391–424. van der Wiel, A. M., and Wijmstra, T. A. (1987a ), Palynology of the lower part (78–120 m) of the core Tenaghi Philippon II, Middle Pleistocene of Macedonia, Greece. Review of Palaeobotany and Palynology 52: 73–88. (1987b), Palynology of 112.8–197.8 m interval of the core Tenaghi Philippon III, Middle Pleistocene of Macedonia. Review of Palaeobotany and Palynology 52: 89–117. van Geel, B., Raspopov, O. M., Renssen, H., van der Plicht, J., Dergachev, V. A., and Meijer, H. A. J. (1999), Quaternary Science Reviews, 18: 331–8. van Zeist, W. and Bottema, S. (1991), Late Quaternary Vegetation of the Near East. Dr Ludwig Reichert, Wiesbaden. Vidal, L., Bickert, T., Wefer, G., and Röhl, U. (2002), Late Miocene stable isotope stratigraphy of SE Atlantic ODP Site 1085: Relation to Messinian events. Marine Geology 180: 71–85. Voelker, A. H. L. and Workshop Participants (2002), Global distribution of millennial-scale records for marine isotope stage (MIS) 3: A database. Quaternary Science Reviews 21: 1185–212. Wade, B. S. and Pälike, H. (2004), Oligocene climate dynamics. Paleoceanography, 19: PA4019, doi: 10/1029/ 2004PA001042. Waelbroeck, C., Labeyrie, L., Michel, E., Duplessy, J. C., McManus, J. F., Lambeck, K., Balbon, E., and Labracherie, M. (2002), Sealevel and deep water temperature changes derived from benthic foraminifera isotopic records. Quaternary Science Reviews 21: 295–305. Watts, W. A. (1988), Europe, in B. Huntley and T. Webb III (eds.), Handbook of Vegetation Science 7. Vegetation History. Kluwer, Dordrecht, 155–92.

137

West, R. G. (1980), Pleistocene forest history in East Anglia. New Phytologist 85: 571–622. Wick, L., Lemcke, G., and Sturm, M. (2003), Evidence of Lateglacial and Holocene climatic change and human impact in eastern Anatolia: high resolution pollen, charcoal, isotopic and geochemical records from the laminated sediments of Lake Van, Turkey. The Holocene 13: 665–75. Willis, K. J. and McElwain, J. C. (2002), The Evolution of Plants. Oxford University Press, Oxford. Rudner, E., and Sümegi, P. (2000), The Full-Glacial forests of central and southeastern Europe. Quaternary Research 53: 203–13. and van Andel, T. H. (2004), Tree or no trees? The environments of central and eastern Europe during the Last Glaciation. Quaternary Science Reviews 23: 2369–87. Wijmstra, T.A. (1969), Palynology of the first 30 metres of a 120 m deep section in northern Greece. Acta Botanica Neerlandica 18: 511–27. and Smit, A. (1976), Palynology of the middle part (30–78 metres) of the 120 m deep section in northern Greece (Macedonia) Acta Botanica Neerlandica 25: 297–312. Wolfe, J. A. (1985), Distribution of major vegetational types during the Tertiary, in E. T. Sundquist and W. S. Broecker (eds.), The Carbon Cycle and Atmospheric CO2 : Natural Variations Archaean to Present. Geophysical Monograph 32, American Geophysical Union, Washington, 357–75. Woodward, F. I., Lomas, M. R., and Kelly, C. K. (2004), Global climate and the distribution of plant biomes. Philosophical Transactions of the Royal Society London B, 359: 1465–76. Woodward, J. C., Macklin, M. G., and Smith, G. R. (2004), Pleistocene glaciation in the mountains of Greece, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations—Extent and Chronology: Part 1. Elsevier, Amsterdam, 155–73. Yasuda, Y., Kitagawa, H., and Nakagawa, T. (2000), The earliest record of major anthropogenic deforestation in the Ghab Valley, northwest Syria: a palynological study. Quaternary International 73/4: 127–36. Yilmaz, Y. (1993), New evidence and model on the evolution of the southeast Anatolian orogen. Geological Society of America Bulletin 105: 251–71. Zachos, J., Pagani, M., Sloan, L., Thomas, E., and Billups, K. (2001), Trends, rhythms and aberrations in global climate 65 Ma to present. Science 292: 68–93.

This chapter should be cited as follows Tzedakis, P. C. (2009), Cenozoic climate and vegetation change, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 89–137.

This page intentionally left blank

5

The Nature and Origin of the Vertebrate Fauna Jacques Blondel

Introduction The aim of this chapter is to provide an account of the complex history of Mediterranean faunas as they evolved from the end of the Pliocene about 1.8 million years ago until the present day. Reconstructing this history is difficult because the Mediterranean basin is one of the most complex regions in the world and is characterized by significant geographical and topographical variation. The Mediterranean basin was formed during the Tertiary by the convergence of the African and Eurasian tectonic plates, in combination with several African microplates, Iberia, and two main African promontories: Apulia in the west and Arabia in the east (Chapter 1 and Dercourt et al. 1986). Where the African and Eurasian plates meet, seismic and volcanic activity have combined with other processes to form a very heterogeneous region. High mountains and deeply dissected topography form the main part of a coastline some 46,000 km in length, 18,000 of which are island shores (Chapter 13). A dominant feature of the region, which has had many consequences for species diversity and the process of differentiation, is the striking contrast between the northern half of the basin with its many large peninsulas—Iberian, Apennine, Balkan, and Anatolian—and the southern half with its more or less rectilinear shorelines. In addition, there is a marked biogeographical contrast between the western and the eastern halves of the Mediterranean, the former having shifted somewhat to the north with respect to the latter (Figure 5.1). The line separating the two north–south ranges in each half of the basin runs approximately along the 36th parallel in the western half and the 33rd in the eastern half. In the western half, west of the

Sicily–Cap Bon line, biota are more boreal in character and overlap to a large degree with those of central Europe. To the east, biota have more affinities with central Asia (Blondel and Aronson 1999). Modern patterns of regional floral and faunal diversity mostly result from differential speciation and extinction rates during the Quaternary (Chapter 4). During the course of the Holocene (the postglacial period that started some 11,500 years ago) the human impact on the terrestrial environment has been of increasing importance (more than in any other region in the world), to the extent that it is sometimes impossible to disentangle the relative contribution of natural from human factors in the design of landscapes and habitats, and hence on the distribution of living organisms. In the final chapter of this book, Blondel and Médail examine the many factors that threaten the plants and animals of the basin in the twenty-first century as well as more general conservation issues (Chapter 23).

The Pliocene-Pleistocene History of Western Eurasia and the Establishment of Mediterranean Biotas As shown in Chapter 4, the Late Tertiary (Late Miocene and Pliocene) was characterized by many climatic changes and these led to important variations of fauna and flora throughout the world. Various lines of research including DNA sequencing and palaeontology have shown that the main lineages of the extant species in temperate and arctic areas of the Northern

140

Jacques Blondel

NE NW 40°N

SW SE

30° 0°

20°

40°E

Fig. 5.1. The four quadrants of the Mediterranean region (modified from Blondel and Aronson 1999).

Hemisphere were generated during the last 10 million years, although many of them are rooted in a more distant past (Lister 2004). For the Mediterranean basin, a large body of data has now been assembled from fields as diverse as palaeobiology, palaeoclimatology, archaeology, molecular phylogenetics, population genetics, macroecology, and evolutionary ecology. These provide important insights into the impacts of global climate fluctuations during the past 2.5 million years or so on living organisms and the distribution of modern faunas. The reconstruction of palaeoenvironments and the identification of major shifts in the distribution of vegetation belts in response to climate changes (e.g. Huntley and Birks 1983) has provided a new perspective on differentiation and extinction rates as well as on the establishment of modern floras and faunas. After the characteristic mediterranean-type climate (with dry and hot summers with intense water stress contrasting with wet and cool winters) became established some 3.2 million years ago, a series of major climatic fluctuations occurred during the Pleistocene with a dominant 100,000-year cycle for the last 1 million years (Webb and Bartlein 1992 and Chapter 4) interrupted by short fluctuations of lower amplitude. These fluctuations occurred at timescales of 103 to 105 years and repeatedly altered the distribution and composition of living systems. Cold conditions prevailed for much of this time so that the Quaternary should really

be viewed as ‘a cold epoch interrupted periodically by catastrophic warm events—the brief interglacials with climate similar to today’ (Davis 1976). Thus, communities such as those of the present day have a short history, of an order of magnitude of a few thousand years at most (McGlone 1996). The repeated shifts in the distribution of the vegetation belts and their associated fauna across Europe in response to Pleistocene climate changes can be considered as migrations sensu stricto, the only difference with the classical annual migration of birds being that the temporal scale of migratory cycles was considerably longer, of an order of magnitude of 100,000 years (Huntley and Webb 1989). These largescale migrations led to the accumulation and concentration of species and communities in the Mediterranean during the ‘glacial winters’ and this had important consequences for biodiversity (Chapter 4). In addition, it has recently been demonstrated that extant elements of the postglacial biota of north-western Europe persisted through glacial periods not only in the ‘classic’ Mediterranean southern refugia, but also farther north in small pockets of favourable microclimates (Bhagwat and Willis 2008). All these movements resulted in gene swamping and mixing, providing few opportunities for allopatric speciation, especially in groups of animals such as birds which disperse over large areas (Blondel 1988). Therefore, as we will see later in this chapter, the Mediterranean includes a surprisingly low proportion of endemic species in many groups such

Nature and Origin of Vertebrate Fauna

141

as birds. Three dominant factors contributed to mould extant faunas in the Mediterranean and these are discussed below.

The Mediterranean Basin as a Refuge During Glacial Times During the most severe phases of the glacial periods, almost no arboreal vegetation survived north of the mountain chains that delimit the Mediterranean region in southern Europe, i.e. the Pyrenees, Alps, and Carpathian mountains. Contrary to what has long been thought, vegetation belts and their associated faunas of the western Palaearctic (the Palaearctic comprises Europe and the non-tropical parts of Asia and Africa) did not simply shift in latitude during glacial times in such a way that Mediterranean biotas would have retired in what is today the Sahara Desert. They actually had to survive within the geographical limits of the Mediterranean basin (Bennett et al. 1991) where they were trapped during glacial episodes because massive west–east barriers, the Sahara Desert and the Mediterranean Sea, prevented their escape south to the tropics (Huntley 1988). Thus, they concentrated around the Mediterranean Sea, especially in the larger peninsulas, Iberian, Apennine, Balkan, and Anatolian, which were larger than they are today because the sea level was about 120 m lower at the global Last Glacial Maximum (Chapter 13). Furthermore, because of the high topographic and climatic heterogeneity of these peninsulas, many refugia existed for plants and animals on mountain slopes and islands as well as in the valleys and gorges of large rivers—even during the most extreme glacial periods (Chapter 4). The resulting mosaic of habitats at the scale of landscapes allowed the coexistence in Mediterranean refugia of assemblages of species from all the vegetation belts of Europe. During glacial times, faunas must have been mixtures of many different communities, without any clear geographic delimitation between Mediterranean and nonMediterranean communities. One example is that of the remains, found in the same deposits referred to as Würm II in southern France, of such species as the Lesser kestrel Falco naumanni (thermophilous Mediterranean species, Figure 5.2), the Ptarmigan Lagopus mutus, the Snowy owl Nyctea scandiaca (birds of the tundra and the alpine belt), and the Rock thrush Monticola saxatilis (saxicolous thrush of Mediterranean mountains) (Mourer-Chauviré 1975). Similar composite faunal assemblages have been found for small mammals and amphibians in fossil deposits of the Riss

Fig. 5.2. The Lesser kestrel Falco naumanni, a typical but declining species found in old cities and craggy areas in the Mediterranean (photo: Otello Badan).

glaciation by Chaline (1972) and Rage (1972) respectively. Such puzzling and disparate assemblages suggest that local landscapes in the Mediterranean during pleniglacial times were kaleidoscopes of such habitats as tundra, steppe, and both coniferous and broad-leaved forests. It is difficult to imagine how these landscapes looked, however, because Late Glacial vegetational communities have no convincing modern analogues (Huntley and Birks 1983). Sea-birds also followed the same general shifts back and forth in relation to glacial–interglacial cycles but few fossil remains allow us to reconstruct their Pleistocene history. One exception is the Great auk Pinguinus impennis (the last individual was shot in 1844 in Iceland). Several Pleistocene and Holocene fossil deposits along the coasts of Iberia, France, and Italy as far east as Calabria show that this large flightless species most probably bred in huge colonies on Mediterranean shores, together with other auks and larids (Mourer-Chauviré and Antunes 1991; Elorza and Sanchez Marco 1993). Holocene remains of Great auk from the Bay of Biscay (the site of Herriko Barra) have been radiocarbon-dated to 5,810 ± 170 years BP, which means that this species (in association with several species of shearwaters, gulls, and alcids) occurred there as a breeding bird during the warmest phase of the Holocene, and was extensively hunted by humans (and sometimes painted by them in caves such as Cosquer cave near Marseilles dating back 20,000 years). These findings raise the question of the breeding range of sea-birds that are today mostly distributed in the northern part of the Atlantic Ocean. It is possible that the breeding range of the Great auk, as well

142

Jacques Blondel

as that of other species such as the Gannet Sula bassana to which it is often associated in fossil sites, may have extended much more to the south in postglacial times than is the case today (Bengtson 1984).

Differential Extinction Processes The second factor refers to differential extinction rates that have been much more severe during glacial times in the western Palaearctic, including the Mediterranean, than in the two other temperate forest blocks of similar size (c.10 million km2 ) in the Northern Hemisphere, namely the eastern Palaearctic and eastern Nearctic. In general terms these three temperate forest blocks shared the same history since the beginning of the Tertiary and are of similar size so that they may be expected to include a similar number of species for any systematic group. This is far from being the case. Latham and Ricklefs (1993) attributed the low taxonomic diversity of tree species in European forests (compared to North America and eastern Asia) to the limited area available for forest taxa during glacial episodes. Presumably, many tree species that occur today in the Nearctic and in eastern Palaearctic became extinct in Europe during Pleistocene glaciations (Jahn 1991 and Chapter 4). If differential extinction of plant species occurred on a larger scale in Europe than in North America, one would also expect more differential extinction in animals. This would explain why European forest-bird faunas are much poorer in species than their Nearctic (Mönkkönen 1994) and east-Palaearctic (Hino 1990) counterparts. Based on the fossil record, Brodkorb (1971) estimated that before the Pleistocene glaciations, the total number of bird species was c.25 per cent higher than today and Mourer-Chauviré (pers. com.) lists at least twenty-one large species (ten Galliforms, four crows, three owls, four bustards) that became extinct in Europe before the end of the Pleistocene. This figure does not include small land birds which are rare in the fossil record. Higher extinction rates in the Palaearctic region were due to the massive east–west barriers of the western Palaearctic, the Sahara, and Arab-Syrian deserts which originated during Late Miocene to Early Pliocene times around 6 million years ago. These barriers prevented large-scale dispersal between North Africa and tropical Africa and western Asia and prevented species from finding refuge in tropical areas. Hence they became trapped north of the Sahara Desert. Large extinction events that resulted from these barriers have been partly compensated for, however, by differentiation processes that occurred

at the population level on the main Mediterranean peninsulas (Hewitt 1996).

The Mediterranean as a Matrix of Differentiation of Populations and Species The third factor refers to differentiation processes of species and populations that have presumably been very active for most groups of plants and animals during glacial times when they were split among several Mediterranean refugia. The legacy of glacial times is the present-day diversity of faunas and floras in the Mediterranean and has been beautifully and repeatedly illustrated by phylogeographic studies. These demonstrate both the differentiation of species at the population level as a consequence of isolation in refugia and their postglacial northward expansion across Europe (Hewitt 2000 and references therein). One example is that of the Brown bear (Ursus arctos) (Figures 5.3 and 5.4). Significant genetic differentiation occurred during glacial times when this species was split into several isolated populations in each of the main Mediterranean peninsulas (Iberian, Italian, and Balkan). Biochemical studies have enabled the construction of intraspecific phylogenies from mitochondrial DNA sequences, which have been used to infer the patterns of European recolonization by these bears as climate improved during the Holocene (Taberlet and Bouvet 1994). Extending these genetic approaches to other groups of animals and plants, Taberlet et al. (1998) have identified several phylogeographic units for each taxon, allowing us to track their migration routes as they recolonized Europe from their Mediterranean refugia. It would be extremely interesting, if it were possible using ancient DNA on a reasonably large sample of individuals, to make similar reconstructions for early humans as climate changed (see Gamble et al. 2004). Phylogeographic studies that combine biogeography and population biology provide valuable insights into the tempo and mode of both the differentiation of species at the population level as a consequence of isolation and their northward expansion across Europe in postglacial times (Hewitt 1996, 2000 and references therein). Although recent studies using molecular systematics based on mtDNA provide evidence that many species of birds are much more ancient than formerly thought and are rooted in the Late Pliocene (Klicka and Zink 1997; Blondel and Mourer-Chauviré 1998), Pleistocene environmental changes certainly had a major

Nature and Origin of Vertebrate Fauna

Pyr Can Nor Dal Abr Slo Cro Gre Bul Rus Est Ro 2 Ro 1 Ursus 12

10 8 6 4 2 Genetic divergence (d, %)

143

Western lineage (lberian refuge)

Western lineage (Balkan refuge)

Eastern lineage americanus

0

Fig. 5.3. Phylogeography of the Brown bear. Isolation of this mammal during glacial periods resulted in the splitting of populations in disjunct refugia. These populations differentiated in three lineages including well-defined haplotypes (Pyr = Pyrenees, Can = Cantabric cordillera, Nor = Norway, Dal = Dalmatia, Abr = Abruzzo, Slo = Slovenia, Cro = Croatia, Gre = Greece, Bul = Bulgaria, Rus = Russia, Est = Estonia, Ro1, Ro2 = Romania). The American bear (Ursus americanus) is used as an out-group, that is a close relative species in which the phylogeny of the brown bear populations is ‘rooted’ (modified from Taberlet and Bouvet in Blondel 1995 and Blondel and Aronson 1999).

influence in shaping extant species diversity. As a large refuge for all the biota of western Europe during glacial times, the Mediterranean basin has been a matrix of differentiation, making it particularly rich in local ecotypes and evolutionary significant units (sensu Moritz 1994; Moritz and Faith 1998). The extant patterns across Europe of the genetic structure of plants and animals (including birds, mammals, amphibians, insects, and humans) is a legacy of this history (e.g. Hewitt 1999, 2000; Taberlet and Bouvet 1994; Taberlet et al. 1998, Kvist et al. 2004 and references therein). Populations from different refugia eventually came into contact, forming hybrid zones and/or contact zones (Taberlet et al. 1998). In some cases, mountain chains acted as biogeographical barriers trapping populations where they differentiated or preventing them from extending northwards. For example, many phylogenetic lineages have been reported to occur only in the Apennine Peninsula because their northward expansion was prevented

by the Alps. Examples are the grasshopper Chorthippus paralellus (Cooper et al. 1995), the wood mouse Apodemus sylvaticus (Michaux et al. 1998), the salamander Salamandra salamandra (Steinfartz et al. 2000), and the Italian hare Lepus corsicanus (Pierpaoli et al. 1999). The fish fauna includes more than 230 species of which 148 are local endemics (Reyjol et al. 2007). It is beyond the scope of this chapter to discuss in detail the history of the fish fauna of the Mediterranean basin, but these animals are especially interesting because their natural dispersal relies entirely on the geomorphological evolution of drainage networks. This is particularly important in the case of the so-called ‘primary freshwater fish species’ which, unlike ‘peripheral fish species’ such as most salmonids, are unable to disperse naturally over extensive areas via the marine environment. The freshwater fish fauna of the Mediterranean basin includes two groups of species: a series of species that are widespread in the whole of Europe, the

144

Jacques Blondel

Fig. 5.4. The Brown bear Ursus arctos still persists in disjunct populations in the Pyrenees, the Apennines, and the Balkans (photo: J.-F. Desmet).

‘Danubian’ species that constitute a rich homogeneous ichyhyofauna in a ‘Danubian district’ (Bianco 1990), and several assemblages specific to more restricted districts around the basin that share few elements in common, either with one another or with the Danubian district.

The Mediterranean Region as a Melting Pot and Crossroads for Animals Background Extant biotas of the Mediterranean are a mixture of elements which either derive from in situ evolution or have colonized the basin from more or less distant regions in various periods in the past. The colonizers constitute the vast majority of present-day species and originated from many biogeographical realms because the Mediterranean, which straddles three continents, Europe, Asia, and Africa, is a crossroads for both plants and animals. For some groups of animals such as mammals, the Mediterranean is really a transitional zone

between different biogeographical realms. Important faunal exchanges and immigrations began during the Late Miocene Messinian Salinity Crisis (Chapter 1) which almost completely desiccated the Mediterranean and partially dried out the Red Sea (6.5–5 Ma) to allow biotic connection between the African and Arabian plates. These connections, which lasted until around 3.5 to 3 Ma, could have allowed the dispersal of east Asian groups to North Africa, as suggested, for example, for mammals by Cheylan (1991), and then to the Iberian Peninsula across the Gibraltar sill during the desiccation of the sea. In the southern part of the basin, the African origin of the Iberian microplate—which collided with Eurasia in the Late Eocene (35 Ma)—explains many African affinities within the flora and fauna of Iberia. Biotic exchanges between Europe and North Africa across the Mediterranean became possible during the Middle Miocene (14–13 Ma) via a continuous landmass separating the ancient sea of Tethys from the Paratethys and roughly corresponding to the Balkan/Anatolian region (Rögl and Steininger 1983). In addition, the land-bridge connection between Sicily and Tunisia that existed before the onset of the Pleistocene presumably

Nature and Origin of Vertebrate Fauna

145

Fig. 5.5. The Ibex Capra ibex whose populations have been dangerously reduced in size is currently expanding in number in many parts of the Alps due to effective conservation and reintroduction programmes (photo: J.-F. Desmet).

also facilitated faunal exchange between Europe and Africa (Voelker 1999).

Large Mammals Perhaps the most spectacular species assemblages of the Late Pleistocene in western Europe and North Africa were those of large mammals to which early humans, both Neanderthal and modern, belonged. Their history is worth summarizing briefly. Mammal faunas of the Mediterranean basin are a legacy of several turnover events and repeated waves of immigration and extinction that go back as far as the Late Oligocene to Early Miocene. Their history is intimately associated with the periodic appearance and disappearance of land-bridges connecting Asia, Europe, and Africa across the Tethys and the Paratethys seas (Chapter 1). In particular, the Levantine corridor (the narrow belt of Mediterranean habitats that stretched along the coast prior to the northward extension of the Sahara Desert in what is now Libya and Egypt) allowed many faunal exchanges

between Africa and Eurasia during the Neogene and the Quaternary (Tchernov 1992). Although the broad geography of the region has stabilized since the Miocene, Pleistocene climatic changes produced many turnover events in the large mammal faunas starting with a progressive decline of the tropical species that entered Europe during the Pliocene. In the Middle Pleistocene, a major faunal turnover associated with large climatic cycles with long-lasting cold episodes was characterized by the disappearance of almost all tropical species, the extinction of many ancient boreal species, the arrival of ‘cold faunas’ of boreal origin, and finally the settling-in of modern faunas that modern humans experienced during the last glacial period (Azzarolli 1983; Bonifay and Brugal 1996, see Table 23.3 in Chapter 23). This fauna included bisons, horses, bears, deers, mammoths, hyenas, panthers, lions, rhinos, reindeers, giant deers, aurochsen, and ibexes (Azzarolli 1983, Figure 5.5). The first true Bison (Bison schoetensacki) replaced Eobison and several new cervids appeared: the Reindeer (Rangifer tarandus), the Red deer

146

Jacques Blondel

(Cervus elaphus) and the Roe deer (Capreolus capreolus). Other mammals of ‘cold origin’ included the Woolly rhino (Coelodonta antiquitatis) and the Muskox (Ovibos pallantis). The primitive boar Sus strozzi was replaced by the modern Wild boar (Sus scrofa), the Mammoth (Mammuthus meridionalis) by a more evolved species (M. trogontherii) which co-occurred with the elephant Palaeoloxodon antiquus. Large carnivores included the modern Wolf (Canis lupus) and the Etruscan bear (Ursus etruscus) which subsequently evolved into the wellknown Cave bear (Ursus spelaeus). In Mediterranean Europe, only a few subtropical taxa (e.g. Hystrix cristata, Macaca sylvanus) from the Early Pleistocene fauna found refuge and survived after the large extinctions of the first glacial stages (Kurtén 1968). The mammal fauna of North Africa evolved in relative isolation but was clearly African in character during the Early and Middle Pleistocene with several species of antelope, an elephant, and many species of rodents. A large turnover of the mammal fauna also took place during the Pleistocene (Kowalski and Rzebik-Kowalska 1991) with savanna-like mammal assemblages of African character, including goats, antelopes, elephants, white rhinos, hares, jerboas, and jackals being enriched by Eurasian species that colonized North Africa using the Levantine corridor during glacial periods of the Late Pleistocene. These Palaearctic elements included the Brown bear (Ursus arctos), Auroch (Bos primigenius), and Deer (Megaloceros algericus, Cervus elaphus). As on the northern shores of the Mediterranean, this period was characterized by an impressive number of large carnivores: Dogs (two species), Lycaon, Fox, Brown bear, Genet, Hyenas (two species), Cats (two species), Lynx (two species), Lion, and Panthers (two species). The main message provided by the mammal history of the Mediterranean is the extraordinary richness and turnover of the large megafauna of this period, a vivid testimony of which is provided by many fossil sites in southern Europe as well as by the superb wall paintings in caves of southern France and northern Spain (e.g. Chauvet et al. 1995). Many of these species became extinct in both Mediterranean Europe and North Africa at the end of the last (Würm) glaciation, probably as a result of a combination of human impact (Martin 1984) and climate changes, as in many other regions of the Northern Hemisphere. The question of the relative contribution of humans—the ‘overkill hypothesis’ (Martin 1984)—and climate change, which led to the transformation of grasslands to forests, to this mass extinction is still debated. But recent studies emphasize the importance of human impact (e.g. Brook and Bowman 2004;

Burney and Flannery 2005). Some survivors of this ancient fauna, which included a surprisingly high number of large predators, disappeared only recently since the lion and elephant survived until historic times in Greece and Syria, respectively, as reported by Herodotus for example. The combination of this profusion of large mammals with the elaborate art of early humans provides a vivid picture of the environment in which people lived in this period. The extant mammal fauna of the Mediterranean includes around 158 non-volant species (Cheylan 1991), with only 25 per cent of these being endemic species. This results from three main factors: 1. the multiple biogeographical origin of species which entered the Mediterranean from Europe, Asia, and Africa; 2. Pliocene and Pleistocene climatic changes that produced periodic faunal turnovers, hence large differentiation diversities over time; 3. severe and ancient (Neolithic and later) human pressures (in the form of habitat changes, animal husbandry, hunting, and the introduction of alien species) that modified natural distribution patterns leading to a decrease in the previously much richer glacial fauna. Because most non-volant mammals are rather weak dispersers and sensitive to geographical barriers, mammalian faunas of the three main sub-regions of the Mediterranean (i.e. Mediterranean Europe, the Middle East, and North Africa) are rather distinct. Most extant components of the European fauna, including the Mediterranean, are of boreal origin (e.g. Sus, Cervus, Ursus) resulting from several immigration waves that brought new faunal elements from temperate Asia during the main glacial episodes. In the Middle East, the mammal fauna is a mixture of species originating from both Eurasia and tropical Africa (Tchernov 1984) as a result of immigration from these two regions during the mid-Pleistocene. In North Africa, Afrotropical elements always dominated the fauna (Jaeger 1975) which was particularly rich in large-hoofed mammals and carnivores, amounting to twenty-nine species before a dramatic human-induced decrease took place during the last few centuries. During this period the three main subregions of the Mediterranean basin experienced massive post-Pleistocene invasions either from the Middle East (e.g. Vormela peregusna, Canis aureus, Microtus guentheri) or from North Africa (Herpestes ichneumon, G. genetta). Other species such as the Rabbit Oryctolagus cuniculus were introduced by humans from

Nature and Origin of Vertebrate Fauna

S

Iron Age 0–500 BC

Neolithic 4,000 BC

L

15 th C. AD

147

L Bronze Age 2,000 BC Neolithic 6,000 BC S

Centre of origin

9,000 BC S

?

L

South-east Asia

Epipalaeolithic 10,000 BC

?

0

500 km

Fig. 5.6. Patterns of invasion of the Mediterranean basin and western Europe by the house mouse. S = sea transportation, L = land invasion (modified from Auffray et al. 1990).

Europe to North Africa or to the Middle East. Two commensal Murid species, the Mouse Mus musculus and the Black Rat Rattus rattus, invaded human settlements from Asia to the Near East at the beginning of the Holocene (Figure 5.6). Finally, several species including the large semi-aquatic rodent Myocastor coypus, the Muskrat Ondatra zibethicus, and the Eastern cottontail Sylvilagus floridanus have recently been introduced to the Mediterranean region from America by humans.

The Mediterranean Bird Fauna The eclectic character of the Mediterranean bird fauna is even more pronounced than that of mammals. The extant breeding bird fauna includes about 345 species compared to the c.500 species that breed in the whole of Europe (Hagemeijer and Blair 1997). After the Late Pliocene extinction in Eurasia of many elements of the bird fauna that occur only in the tropics today—such as Cracidae, Psittacidae, Musophagidae, Coliidae, Trogonidae, and Bucerotidae (Brodkorb 1971; Bochenski 1985; Blondel and Mourer-Chauviré 1998)—the Pleistocene-Holocene development of avifaunas in the Mediterranean involved faunal elements from nine

biogeographic units (see Blondel 1988 and Figure 5.7). Two of these units played particularly important roles: Eurasia (150 species) and the semi-arid belts of the south and south-east margins of the region (85 species). In contrast to mammals, the bird fauna of the southern Mediterranean margins has close affinities with that of the Asiatic steppes as a result of an active radiation within the wide belt of semiarid habitats that encircles the western Palaearctic from the Atlantic Ocean to the Arabian plate and the steppes of south-central Asia. Very few species originated in tropical Africa, one of them being the Fan-tailed warbler Cisticola juncidis (Figure 5.8). Most Mediterranean birds belong to four broad ecological categories: steppe species (88 species), forest species (76), water-birds (69), and shrubland species (41). Although we should expect the last to be numerous and dominant in the many types of matorrals that extend over more than half the region, this category is poorly represented (12%), mostly because many Mediterranean matorrals are of secondary, anthropogenic origin (Blondel and Aronson 1999). As explained in the first section of this chapter, the striking telescoping of the European faunas within the Mediterranean basin during cold glacial times prevented geographical isolation of species that would have been

148

Jacques Blondel

123

17

10

8

47

77

17

16 14

14

Fig. 5.7. Biogeographical origin of the bird fauna of the Mediterranean region (delimited by the shaded area). Pictograms around the figure represent the Old World biogeographical origins of the species. The numbers therein are the numbers of species originating in each area with forty-seven species of Mediterranean origin and thirty-six (asterisks) being regional endemics (modified from Blondel 1988).

a prerequisite for allopatric speciation. This explains both the very low number of endemic bird species in Mediterranean forests (three Nuthatches, one Tit, and two Pigeons, the last in the Canary Islands) and the homogeneity of forest-bird faunas everywhere in Europe. Different types of more or less isolated patches of matorral developed in the Mediterranean region as early as the beginning of the Pleistocene (Suc 1984) and persisted throughout the Quaternary in local areas where climatic, edaphic, and topographical conditions allowed the existence of only a shrubby vegetation (Pons 1981). Examples of species that evolved in these patches of matorral are warblers of the genus Sylvia (Figures 5.9 and 5.10), Partridges of the genus Alectoris (Figure 5.11), and some others such as the two Rock thrushes Monticola spp., and the Moussier’s redstart Phoenicurus moussieri, which is endemic in North Africa.

Invertebrates A detailed discussion of the origins, distribution, and patterns of differentiation of most invertebrates is beyond the scope of this chapter. Many groups of invertebrates have disjunct distributions, e.g. between the western and eastern parts of the basin, with a high number of endemic species in several areas (Sanmartin 2003). These distribution patterns result from the complex palaeogeographical history of the region and can be explained either by fragmentation of widespread ancestors through vicariant (isolating) events, or by dispersal across pre-existing barriers. Reconstructing vicariant events is complicated by the fact that the biogeographical history of the Mediterranean basically results from the accretion through time of multiple land masses. Barriers to dispersal such as seaways or mountain chains repeatedly appeared and disappeared

Nature and Origin of Vertebrate Fauna

149

Fig. 5.8. The Fan-tailed warbler Cisticola juncidis, one of the few species of tropical origin which colonized the Mediterranean region from tropical Africa where no less than forty species live in savanna-like habitats (photo: Otello Badan).

through time, resulting in a pattern of ‘reticulate’ area relationships, in which repeated episodes of vicariance and dispersal have affected the same areas at different times (Sanmartin et al. 2001). Mosaics of moving island arcs, basins, and fragmenting tectonic belts have resulted in a ‘reticulate’ biogeographical history, in which Mediterranean biota repeatedly fragmented and merged as dispersal barriers appeared and disappeared through time (Oosterbroek and Arntzen 1992). Again, detailed discussion of the reconstruction of their biogeography is beyond the scope of this chapter but new methods of ‘dispersal-vicariance’ analysis including both dispersal and vicariance scenarios (Ronquist 2002) have been used in a study of the biogeographical history of a group of beetles (the subfamily Pachydeminae, Scarabaeoidea, Coleoptera). This group with low dispersal abilities is widespread in the southern Palaearctic with many disjunct distributions, reduced geographical ranges, and many local endemic species. It is a good example of how many groups of insects become established in the Mediterranean where both vicariance and

dispersal events contributed to shape the extant biogeographical patterns in the region.

Landscape Structure, Habitat Heterogeneity, and Scales of Diversity All the areas and ecological systems that stretch from the top of the mountain ranges that encircle the basin down to the sea can be considered as experiencing a Mediterranean-type bioclimate (Chapter 3). Clockwise around the basin from south-west Europe the main mountain chains are the Pyrenees, the Alps, the Dinaric Alps, the Balkan mountains, Pontic and Taurus mountains, the Caucasus, the Zagros mountains, the Lebanon ridges, and the Atlas and Anti-Atlas ranges in North Africa. Except for the long strip of lowland desert that meets the sea between Tunisia and the Sinai, one is almost never out of sight of mountains and slopes steeper than 20 per cent occur in more than

150

Jacques Blondel

Fig. 5.9. Four species of warbler that are typical of Mediterranean matorrals where they evolved: the Mediterranean warbler Sylvia melanocephala (a), the Subalpine warbler S. cantillans (b), the Dartford warbler S. undata (c), and the Marmora’s warbler S. sarda (d) which is endemic on islands of the western part of the basin (photos: Jacques Blondel).

half of the territory (McNeill 1992). High mountain chains also occur within the basin with examples being the Apennines, the many cordilleras in Spain, the Middle Atlas, the Rif mountains (Morocco), the Tellian Atlas (Algeria), and the mountains of Greece (Chapter 1). In addition, most islands of the Mediterranean are themselves mountains in the sea so that an alternative name for the Mediterranean might be the ‘Sea-among-the-Mountains’ (McNeil 1992; Blondel and Aronson 1999). Some of these mountains exceed 4,000 m and completely isolate upland basins or plateaux, as well as a myriad of valleys opening towards the sea. A series of eight vegetation belts or ‘life-zones’ along these altitudinal and latitudinal gradients has been determined by Quézel (1985). Ranging from the

‘infra-Mediterranea’ with a near-desert climate in North Africa to the ‘cryo-Mediterranean’ at the highest elevation, they determine how biodiversity is distributed along habitat gradients (Figure 5.12). The high topographical and climatic heterogeneity, across which human action has superimposed a wide variety of land uses, is perhaps the most notable feature of the Mediterranean. The resulting kaleidoscope of habitats and ecosystems includes snow-capped mountains, barren hills, a myriad of matorral types, lush forests and glades, craggy karst ravines, sandy beaches, islands and islets, river channel networks, and wetlands. All these environments are discussed in detail in Part II of this volume. This variety of habitats is associated with a wide gradient of climates ranging from almost subtropical to

Nature and Origin of Vertebrate Fauna

Fig. 5.10. The Rock partridge Alectoris graeca, one of the four endemic Partridge species of the Mediterranean region (photo: J.-F. Desmet).

alpine conditions (Chapters 3 and 12). Even within the same country, for example Greece, farmers in Macedonia and Epirus must feed their livestock on stored hay in the winter because there is nothing to graze, while those in the Cyclades and Crete have to feed their livestock at the height of summer when the land is completely dry and without any green plants (Catsadorakis 2003). The highest diversity of both plant and animal species does not necessarily occur in mature forests but quite often in open areas on nutrient-poor soils, usually after a disturbance event such as fire which triggers ecological successions (Chapter 7). These features show that differentiation diversity (species turnover or ‘beta’ diversity) which refers to compositional change along habitat or geographical gradients is particularly high as a result of habitat specialization and vicariance. This does not mean, however, that regional diversity is necessarily higher than further north in Europe for all groups of animals. Although regional diversity (at a scale of 102 to 106 m2 ) ranks among the highest in the world for plants, this is not the case for all taxonomic groups. In birds, for example, the highest level of regional diversity along latitudinal gradients does not occur in the southern part of Europe as expected from the well-known north–south gradient of species richness, but further north in the nemoral belt of deciduous and mixed forests (Gregory et al. 1998). In fact, a pattern of maximum richness in mid-latitudes, rather than in the south, has already been demonstrated for North American birds by Mönkkönen (1994) and for European birds by Blondel and Farré (1988). However, differentiation diversity is still much higher in the Mediterranean than in any other part of Europe as a result of the combination of local (alpha) diversity and differentiation diversity in relation

151

to the exceptionally high environmental heterogeneity of the region (Cowling et al. 1996). This heterogeneity may be exemplified by the huge variety of shrublands that range from the high maquis covered by a cluster of evergreen leathery-leafed, deep-rooting plants such as holly oak, strawberry tree, mastic tree, mock privet, holm oak, carob, turpentine tree, heather, myrtle, wild olive, oleander, Phoenician and other junipers, to the dry Greek phrygana with its thin hairy, spiny, rounded plants that hug the ground (Chapter 7). The runoff that drains from the Mediterranean mountains converges in the lowlands in networks of stream channels and larger river systems that flow across large alluvial plains and deltas. The latter environments commonly include a mixture of wetlands, salt marshes, forests, and grasslands. Examples are the Guadalquivir, Ebro, Rhône, Pô, Evros, and Nestos rivers and their catchments. The Mediterranean region includes about 30,000 km2 of wetlands (Chapter 9) which are among the most productive and diverse ecosystems of the basin as well as famous breeding areas for many species of terns, waders (Figure 5.13), herons, and ducks and stopover sites for thousands of migrating geese, ducks, and waders. In addition, each of the 5,000 islands and islets of the basin shows an almost unique array of physical, bioclimatic, and geobotanical features. Most of them have been entirely disconnected from any continent since at least the Messinian Salinity Crisis (Chapter 1), contributing to local differentiation of biotas which include many endemic species. The natural structure of landscapes and habitats makes the whole Mediterranean region a huge mosaic of habitat patches that produces an enormous variation of local selection regimes and adaptations. The resulting phenotypic plasticity or local specialization, which depends on the tension between gene swamping and the strength and direction of local selection pressures, is one of the most fascinating albeit still poorly known aspects of Mediterranean wildlife. Few empirical studies have examined in detail how proximate and ultimate factors determine phenotypic variation, that is the variation of morphological or any other life-history trait, in relation to small-scale environmental variation. This question has been addressed from a long-term monitoring programme in the Mediterranean region using Blue tit (Cyanistes caeruleus, Figure 5.14) populations living in habitat patches dominated either by deciduous (Quercus humilis) or evergreen (Q. ilex) oaks, which strongly differ in the phenology and abundance of the invertebrates associated with them and that constitute the main food for tits (Blondel et al. 2006). This study has shown that phenotypic variation and local differentiation of

152

Jacques Blondel

S. sarda S. undata S. deserticola S. melanothorax S. cantillans S. rueppelli S. melanocephala S. mystacea S. conspicillata S. communis S. nisoria S. curruca S. leucomelaena S. hortensis S. minula S. nana

Western Mediterranean Central Mediterranean SB Eastern Mediterranean

SCs SNi SNi

Central

SU

Western

SCu SCu

SA

SS

SH

SM

SMy SMy SR

SD

SC

SMi SMx

SCo

0

500 km

SL

Eastern

Fig. 5.11. Relationships between the three main groups of Mediterranean warblers (genus Sylvia) and their geographical range. Symbols refer to the ‘centre of gravity’ of each species, that is the middle of their distribution range. Differentiation within these three groups is supposed to have occurred as a result of repeated episodes of isolation and re-expansion of matorral-like habitats in lowland areas of these three regions during the Pleistocene. SU = S. undata, SS = S. sarda, SD = S. deserticola, SC = S. cantillans, Sco = S. conspicillata, SM = S. melanocephala, SR = S. rueppelli, SMx = S. melanothorax, SH = S. hortensis, SL = S. leucomelaena, Smy = S. mystacea, Smi = S. minula, Scu = S. curruca. The ‘centres of gravity’ of the four mid-European species (SA = Sylvia atricapilla, SB = S. borin, SNi = S. nisoria, SCO: S. communis) are also shown. The dashed lines in the phylogenetic tree refer to relationships that involve some uncertainty (modified from Blondel et al. 1996).

fitness-related life-history traits at the scale of habitat mosaics can result either from genetically based local specialization, or from adaptive phenotypic plasticity, which allows the same genotype to cope with various environmental conditions, depending on the direction and strength of selection pressures as well as on dispersal abilities (e.g. Blondel et al. 1999, 2001, 2006). This is a good example of the myriad of local ecotypes that result from environmental heterogeneity in the Mediterranean. Recent work has investigated the impact of climate forcing on Blue tit populations and

the significance of summer climatic conditions for adult survival (Grosbois et al. 2006). Another kind of turnover rate that occurs in time is the huge number of birds that use the Mediterranean either as a stopover in the course of their migrations or for spending the winter. The Mediterranean is a key area for wintering birds, especially ducks and geese, but also many species of waders, raptors, and passerines (Figures 5.15 and 5.16). Long-distance trans-Saharan migrants that have to cross the Mediterranean Sea and then the Sahara Desert to reach their African winter

Nature and Origin of Vertebrate Fauna

153

4000 a b

Veg etat ion line

Altitude (m)

3000 c

m»

Tree li

ne

d 2000

m » –5 °C m » –2 °C m » +1 °C

e f 1000

m » +3°C

g h

0 30

–9°C

m » +7

m » +5°C

°C

34

38 Latitude (°N)

42

46

a: Cryo-Mediterranean b: Alti-Mediterranean c: Oro-Mediterranean d: Montane-Mediterranean e: Supra-Mediterranean f: Meso-Mediterranean g: Thermo-Mediterranean h: Infra-Mediterranean

Fig. 5.12. The zonation of the various vegetation belts in the western Mediterranean area in relation to both altitude and latitude. m = the average of the minima of the coldest month (modified from Le Houérou in Blondel and Aronson 1999).

Fig. 5.13. Examples of nest types for terns and waders that can be found on small islets within Mediterranean lagoons between April and July. Nests of Avocet Recurvirostra avosetta (a), Gull-billed tern Gelochelidon nilotica (b), Little tern Sterna albifrons (c), and the Common tern S. hirundo (d) ( photos: Jacques Blondel).

154

Jacques Blondel

Fig. 5.14. The Blue tit (Cyanistes caeruleus) which has been successful in colonizing most forested or semi-forested parts of the Mediterranean region as far south as the Atlas Mountains and the Canary Islands (photo: P. Perret).

quarters have to find in the Mediterranean the resources they use as fuel, especially berries, which are rich in easily metabolizable carbohydrates. It has been estimated that 5 billion birds cross the Mediterranean basin each autumn (Berthold 2001).

Patterns of Endemism in Mediterranean Vertebrates One may expect high endemism rates in a region characterized by an exceptionally high degree of geographic isolation of many areas. The number of rare and endemic taxa that survive as small populations is indeed quite high in plant species since around half the 23,000 plants of the Mediterranean are endemic (Quézel 1985). Endemism rates on islands are particularly high in many groups, especially plants with an average of 20 per cent of the taxa. Before their colonization by humans, most of the larger Mediterranean islands were populated by odd assemblages of animals including tortoises, giant rodents, flightless owls, and dwarf deer, hippos, and elephants. Many of these endemic animals evolved striking adaptations such as gigantism or dwarfism, as exemplified by the now extinct extraordinary assemblages of mammals which included dwarf hippos and elephants in many of the larger Mediterranean islands (Simmons 1988; Diamond 1992; Blondel and Vigne 1993, Figure 5.17).

No fewer than twelve species of dwarf descendants of the ancestor elephant Palaeoloxodon antiquus inhabited Mediterranean islands (Lister 1996) and even the small island of Tilos (64 km2 ) in the Aegean Sea had its own species of dwarf elephant. There is good evidence that all the endemic species of mammals, except three shrews (Crocidura sp.) and a rodent (Mastomys sp.) in Crete, have been directly or indirectly eradicated by humans who established permanent populations on the larger islands from around 9,000 years ago (Simmons 1988) and even earlier. According to Davis (2003), Cyprus was uninhabited by humans before the 9th millennium BC and settlement of the island may have been as a result of a demographic pressure on the nearby mainland. Although we still lack clear evidence for the coexistence of humans and the endemic Cypriot hippo and elephant fauna, Simmons (1988) proposed that the hippo and elephant bones found with cultural remains at Akrotiri, Cyprus, indicated a pre-Neolithic kill-site. It could also be the case that the hippo/elephant fauna was supplanted by the new imported mainland fauna, which included close competitors such as pigs that could have formed wild feral populations after they escaped from human control. All species of the extant mammal fauna of Mediterranean islands are derived from animals that were originally introduced by humans either as domestic species (sheep, goat, pig, cattle, dog) or as stowaways on boats. Several of them such as the Mouflon and the pig formed feral populations that still occur today. Large carnivorous species such as the wolf, the wild cat, and the lynx never immigrated, probably because they are potential competitors with humans (Poplin 1979). The introduction by humans of all the modern non-volant mammals on Mediterranean islands resulted in a very disharmonic fauna and the complete post-Pleistocene turnover of species (Figure 5.18). Endemism rates differ greatly according to groups, mostly as a result of taxon-specific variation in dispersal ability (Table 5.1). Endemism rates are especially high in less mobile animals such as insects, reptiles (Figure 5.19) and amphibians, and of course, in fish species that are isolated in the large number of local catchments and large lakes throughout the region (Chapter 9). A myriad of springs, small rivers, deep ravines, lakes, ponds, and subterranean caves form a complex network of water bodies that has allowed for the differentiation of many populations and species. With 148 endemic species and a large number of subspecies in thirteen families, the freshwater fish fauna of the northern part of the Mediterranean is much more diversified than that of temperate Europe further north (Crivelli

Nature and Origin of Vertebrate Fauna

155

Overwintering

Autumn migration

O S A

N D J

J J M

F M A

Spring migration

Breeding Fig. 5.15. The Mediterranean as a key place for migratory and wintering birds. Most migrating species of the Palaearctic region either migrate through the Mediterranean basin to reach their African winter quarters or overwinter in the basin. A few species overwinter in tropical regions of southern Asia. The figure also shows (bottom right) the turnover through seasons of breeding, migrating, and wintering birds. Note the overwhelming importance of autumn migrants and wintering birds, and the length of the breeding season which largely overlaps with the spring migration (modified from Blondel and Aronson 1999).

and Maitland 1995; Figure 5.20). The families with the more speciose species are the Cyprinidae (81 species), followed by Cobitidae (14), Gobiidae (10), Cyprinodontidae (6), and then Salmonidae (8). Fish species richness is mostly determined by contemporary climate regimes but historical biogeography, especially Pleistocene glaciations, played an important role in producing differential extinction events so that there is a north–south gradient of increasing fish diversity across Europe (Oberdorff et al. 1999). Most endemic fish species live in small rivers, streams, and lakes and sometimes in the upper part of the river catchments. The richest region in fish freshwater species is Greece with 106 species that differ-

entiated in the many rivers and streams of the Pindus Mountains. Interestingly, many of the highest peaks in the Pindus Mountains were glaciated during cold stages of the Pleistocene (Chapter 12). Endemism rates are much lower in groups with larger distributions and dispersal ranges such as mammals and especially birds, which explains why few species of larger vertebrates are endemic in Mediterranean forests. No more than 18 per cent of the endemic bird species in the Mediterranean are forest birds, e.g. the three Mesogean Nuthatches (Sitta whiteheadi, S. krueperi, S. ledanti), one tit (Parus lugubris), one woodpecker (Picoides syriacus), and two Pigeons in the Canary Islands

156

Jacques Blondel

Fig. 5.16. The Short-toed eagle Circaetus gallicus, a summer breeding visitor that also migrates in huge numbers through the Mediterranean region. This eagle, which feeds almost exclusively on snakes, could be under threat in some Euro-Mediterranean countries by the widespread recovery of forests (photo: Jacques Blondel).

(Columba trocaz, C. junoniae) (Blondel 1988). Among rare endemic species, the Spanish Imperial eagle (Aquila adalberti) is one of the most threatened bird species of the Mediterranean with no more than 150 or so breeding pairs left in forests of southern Spain. The species has shown some slight recovery in recent years, presumably as a result of active protection (Ferrer and Donazar 1996). In birds, most endemic taxa are species of open or semi-open shrubby habitats (matorrals). For example, Blondel and Farré (1988) have shown that as vegetation develops in the course of successional processes, bird species of Mediterranean origin are progressively replaced by species of mature forests, which are widespread everywhere in Europe. Endemic mammal species are often more flexible in habitat selection except the very few endemic species in Mediterranean islands. Examples of endemic taxa are the Barbary macaque (Macaca sylvanus) in northern Africa, several species or subspecies of Ibex (Capra ibex) (the Libyan and the Spanish Ibex), the Bezoar Goat (Capra aegagrus) in

Anatolia, and the Iberian Lynx (Lynx pardellus), a very rare species with probably no more than 100 individuals left, in southern Spain. Among the oddities of the Mediterranean basin, the subterranean karstic fauna of the Balkans is the richest in the world with more than 500 species of stygobionts (mostly cnidarians, freshwater sponges, tubeworms, and clams) and 780 species of troglobionts (Sket 1999). Most of these species are local endemics of unique scientific interest, quite often in a single cave. These animals are sightless and colourless, with highly developed tactile organs that enable them to survive in a world where touch is the primary sense. Their low fecundity and high adult survival make them among the best examples of so-called demographic K -strategists. Perhaps the most emblematic of these stygobiont animals is the blind Cave salamander Proteus anguinus, an almost eyeless and pigmentless salamander-like animal, 25 to 50 cm in length, whose distribution is restricted to the caves and

Nature and Origin of Vertebrate Fauna

Fig. 5.17. Pygmy hippos and elephants (roughly the size of pigs and dogs), which were widespread on some of the larger Mediterranean islands at the time when humans first arrived on them some 10,000 years ago. Note the tiny size of the mammals relative to that of the land turtle and swans (modified from Blondel 1995 in Blondel and Aronson 1999).

underground rivers of the Dinaric karsts from northeastern Italy to southern Herzegovina. Proteus lives in slow-running, underground rivers and in deep caves (Hervant et al. 2000). As in all stygobionts its metabolic rate is very low, which is an adaptation to environments where organic matter and especially animal prey are particularly scarce in this lightless habitat. Local biodiversity per cave is usually rather low due to high fragmentation and isolation of habitats but as a result of isolation for millions of years, species differ much between caves that are only a few kilometres apart. Mediterranean karst environments are discussed in Chapter 10.

Human Activity and Land Use Change in the Holocene: The Impact on Biodiversity The thousands of years of human activity in the Mediterranean have resulted in both increases and

157

decreases in biodiversity in different regions and at different times. Although human impact in the past as well as the present has had many detrimental effects on biological diversity, this has certainly not been the case in all parts of the region at all times (Blondel 2006; 2008). The main factors include the physical geography of the region, the socio-economic status of societies, and the land use practices. Perhaps the most intriguing event that occurred at the end of the Pleistocene was the collapse of the large herbivores, which has been attributed, at least partially and perhaps erroneously (see Owen-Smith 1987; Brook and Bowman 2004; Burney and Flannery 2005), to hunting by Upper Palaeolithic humans. It is important to appreciate, however, that human activity is not always detrimental to biodiversity. In some cases, a kind of coevolution between humans and natural systems presumably made them quite robust against disturbance. Some authors (e.g. di Castri 1981), argue that these narrow relationships between humans and plants and animals have a complex ‘co-evolution’ that has shaped the interactions between Mediterranean ecosystems and humans through long-lasting but constantly evolving land use practices (Chapter 7). This may explain why Mediterranean ecosystems are less vulnerable to invasive species than other Mediterranean-type ecosystems such as those of California, Chile, South Africa, and Australia (Drake et al. 1989; Chapter 23). Over the past ten millennia or so, biological diversity in the Mediterranean basin has depended on the many interactions between humans and other living systems. Human activities have had positive as well as negative effects on biodiversity so that the effect of human action on Mediterranean biodiversity cannot be assessed without a careful analysis of its positive and negative aspects. Large mammals paid a heavy cost to human persecution towards the end of the last glacial stage to the extent that the present-day mammal fauna of the region is but a weak reflection of what it was in Late Pleistocene times. However, in contrast, few extinction events appear to have occurred for most groups of Mediterranean plants and animals during the millennia of traditional rural life styles, until the end of the nineteenth century (Chapter 23). On the whole, the impact of humans and domestic animals on vegetation and ecosystems resulted in a clearing of the native vegetation and cutting of trees and shrubs for fuel and forage. A moderate clearing and lowering of vegetation height undoubtedly favoured biodiversity by adding new habitats within landscapes that were primarily forested (Blondel and Aronson 1999). It is only when human population density reached some threshold that

Jacques Blondel

Uncalibrated time (×1000 years)

Ho m O v o sa is a pie Ov rie ns is s Cap a. m u Su ra hir simo s s cu s n Su . do s m Vu s. sc esti lpe rof cu Can s vu a s Bo is famlpes s Eri tauru iliaris n Ap aceo s o u Gli demu s eur s g s s op l Mu is ylv aeu atic s s Su m. d us n cu o m Cro s e s t c e Elio iduratrusc icus my su us Mu s q ave s Equ tela uercin olens n Equus ca ivalis us b u Fel s as allus is s inu Fel . l. s is c Cer s. l. atus vus rey Ra i t ela Urstus ra phus Lepus ar ttus c R a u s co t o s ttu rsi Ory s no can u c r ? M tolag vegic s art us c us es ma unicu rte lus s

158

0 Middle Ages -1 Antiquity -2 Bronze and -3 Iron Ages -4 Neolithic

-5 -6 -7

Mesolithic -8

Cyn

ot Me heriu Epi cace m sor ro s Rh uculu Tyr agam s rhe ys n Pro icola lag us

-9

Fig. 5.18. Turnover of non-volant mammal species in Corsica as a result of human colonization around 9,500 years ago. Vertical lines indicate duration of presence in each case. The highly endemic mammal fauna, which included six species at the end of glacial times, has been completely replaced by twenty-five species (including humans) which progressively entered the island during the course of the Holocene (modified from Blondel and Vigne 1993).

biodiversity declined, mostly as a result of overgrazing, which produced a drastic reduction of plant cover, biomass, and productivity, lower fertility and enhanced runoff, erosion, and sedimentation (Le Houérou 1990) (Chapter 20). The large-scale forest clearances were obviously detrimental to several species and communities but the new landscapes designed and repeatedly redesigned by humans in the course of centuries have provided opportunities for many species to colonize new habitats. For plant and animal species, the main consequences of traditional human-induced changes in Mediterranean habitats have not been so much a decrease of overall species richness at a regional scale, than a tremendous advantage for species adapted to drylands and shrublands, as opposed to forestdwelling species. Indeed, human impact sometimes resulted in a significant increase in biodiversity. For many centuries, traditional land use systems such as

the Roman triad Sylva-saltus-ager (forest-rangelandcrop fields) or the Iberian dehesa-montado systems resulted in changing forest landscapes in mosaics of TABLE 5.1. Endemism rates of several groups of species in the Mediterranean Group Vascular plants Freshwater fish1 Reptiles1 Amphibians1 Mammals Birds Insects Butterflies

No. of species 25, 000 300 165 63 197 366 150, 000 321

Endemism (%)

World no. of species (%)

50 44 68.5 58.7 25 17 — 46

10 ? 2.5 1.5 4.2 3.8 0.6 ?

Note: The figure given for the number of species of insects is probably a significant underestimation. 1

For the northern part of the Mediterranean basin.

Source: Blondel and Aronson 1999.

Nature and Origin of Vertebrate Fauna

159

Fig. 5.19. The very large and beautiful Eyed lizard Lacerta lepida, an endemic species of the Iberian Peninsula, southern France, and some parts of North Africa (photo: Jacques Blondel).

habitat patches providing different resources for people. Such a design of landscapes represented a golden age for Mediterranean landscapes and ecosystems resulting in an increase of several components of diversity,

mostly the beta-diversity (species turnover) and the gamma diversity (total number of species occurring in all habitat patches at the landscape scale) (Blondel et al. 2006). For example, in a recent archaeological

Balkans

Iberian Turkey

Italian

Cyprinidae Cobitidae Cyprinodontidae Gobiidae Salmonidae others

0

500 km

Fig. 5.20. Levels of endemism of freshwater fish of the large peninsulas of the northern part of the Mediterranean basin (modified from Crivelli and Maitland 1995, in Blondel and Aronson 1999).

160

Jacques Blondel

study of Roman land use practices in France, Dupouey et al. (2002) demonstrated that ancient cultivation by the Romans added a significant amount of plant diversity to the forest. Of a total of 116 species recorded in a forest block, 29 are only present in or near the ancient Roman houses, enclosures, and nearby terraces, whereas 15 species are only present in remote terraces and undisturbed areas. Some anthropophilous species such as the Periwinkle Vinca minor or the gooseberry bush Ribes uva-crispa are characteristic of Roman ruins. Speciation events may have been enhanced by the continuous redesign of landscapes and habitats by humans. Such redesign may have resulted in vicariance events causing allopatric distributions in which populations become isolated from one another. Small founder populations and subsequent genetic changes that they experienced in their new habitat conditions could have led to the formation of new species. Although many hybridization events fail to produce fertile taxa, some do so and some of them, following changes in chromosome arrangements, may form new genetically isolated and distinct species (Abbott 1992). The combination of local differentiation during glacial times—in the myriad of separate catchments—and human-induced selection made many local ecotypes an ‘option value’ because of their potential to provide an economic benefit in the future. For example, the Mediterranean basin includes 145 varieties of domesticated bovids and 49 varieties of sheep (Georgoudis 1995). Over the centuries, hundreds of varieties of olive, almond, wheat, and grape of economic value—which have been selected by humans— definitely added to the biological diversity of the Mediterranean (Blondel and Aronson 1999). As pointed out by Diamond (2002), the human influence on populations through the process of domestication undoubtedly constituted an important selective factor in their evolution. As an example, the olive tree—which is the most emblematic plant species of all Mediterranean cultures— currently constitutes a complex of many wild forms (Olea oleaster) as well as weedy types classified as O. europaea var. sylvestris, and many cultivars classed as O. europaea var. europaea (Chapter 7). Recent genetic studies have shown that selection on cultivars has occurred in different genetic pools, showing that olive domestication occurred in many parts of the Mediterranean basin as far back as the fourth millennium BC in Palestine (Besnard et al. 2002). Overall, it appears that traditional human activities have in fact been beneficial for many components of biological diversity in the basin. Gomez-Campo (1985) and Seligman and Perevolotsky (1994) among others

have argued that the highest species diversities in the Mediterranean area are found in frequently but moderately disturbed sites, giving support to the diversitydisturbance hypothesis (Chapter 7). The collapse of most traditional land use systems at the end of the nineteenth century and especially after the Second World War completely changed land use patterns and resulted in new systems of land management that often raise serious problems of conservation; these themes are explored more fully in the final chapter of this volume (Chapter 23).

References Abbott, R. J. (1992), Plant invasions, interspecific hybridization and the evolution of new plant taxa. Trends in Ecology and Evolution 7: 401–5. Attenborough, D. (1987), The First Eden. The Mediterranean World and Man. Fontana/Collins, London. Azzaroli, A. (1983), Quaternary mammals and the ‘endVillafranchian’ dispersal event—a turning point in the history of Eurasia. Palaeogeography, Palaeoclimatology, Palaeoecology 44: 117–39. Bengtson, S. A. (1984), Breeding ecology and extinction of the Great Auk (Pinguinus impennis): Anectodal evidence and conjectures. The Auk 101: 1–12. Bennett, K. D., Tzedakis, P. C., and Willis, K. J. (1991), Cold stage refugia of north European trees. Journal of Biogeography 18: 103–15. Berthold, P. (2001), Bird Migration: A General Survey (2nd edn.) Oxford University Press, Oxford. Besnard, G., Khadari, B., Baradat, P., and Bervillé, A. (2002), Olea europaea (Oleaceae) phylogeography based on chloroplast DNA polymorphism. Theoretical and Applied. Genetics 104: 1353–61. Bhagwat, S. A. and Willis, K. J. (2008), Species persistence in northerly glacial refugia in Europe: a matter of chance or biogeographical traits? Journal of Biogeography 35: 464–82. Bianco, P. G. (1990), Potential role of the palaeohistory of the Mediterranean and parathetys basins on the early dispersal of Euro-Mediterranean freshwater fishes. Ichthyological Exploration of Freshwaters 1: 167–84. Blondel, J. (1988), Biogéographie évolutive à différentes échelles: l’histoire des avifaunes méditerranéennes, in H. Ouellet (ed.), Congressus Internationalis Ornithologici. National Museum of Natural Sciences, Ottawa, i. 155–88. (1995), Biogéographie. Approche Ecologique et Evolutive. Masson, Paris. (2006), Man as ‘designer’ of Mediterranean landscapes: a millennial story of humans and ecological systems during the historic period. Human Ecology 34: 713–29. (2008), On Humans and wildlife in Mediterranean islands. Journal of Biogeography 35: 509–18. and Aronson, J. (1999), Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford. and Farré, H. (1988), The convergent trajectories of bird communities in European forests. Oecologia 75: 83–93. and Mourer-Chauviré, C. (1998), Evolution and History of the western Palaearctic avifauna. Trends in Ecology and Evolution 13: 488–92.

Nature and Origin of Vertebrate Fauna Blondel, J. and Vigne, J.-D. (1993), Space, time, and man as determinants of diversity of birds and mammals in the Mediterranean region, in R. E. Ricklefs and D. Schluter (eds.), Species Diversity in Ecological Communities. Chicago University Press, Chicago. 135–46. Catzeflis, F., and Perret, P. (1996), Molecular phylogeny and the historical biogeography of the warblers of the genus Sylvia (Aves). Journal of Evolutionary Biology 9: 871–91. Dias, P. C., Perret, P., Maistre, M., and Lambrechts, M. M. (1999), Selection-based biodiversity at a small spatial scale in an insular bird. Science 285: 1399–402. Perret, P., Dias, P. C., and Lambrechts, M. M. (2001), Is phenotypic variation of Blue tits (Parus caeruleus L.) in Mediterranean mainland and insular landscapes adaptive? Genetics Selection Evolution 33 (suppl. 1): S121– S139. Thomas, D. W., Charmantier, A., Perret, P., Bourgault, P., and Lambrechts, M. M. (2006), A thirty-year study of phenotypic and genetic variation of blue tits in Mediterranean habitat mosaics. BioScience 56: 661–73. Bochenski, Z. (1985), The development of western Palaearctic avifaunas from fossil evidence, in Proc. XVIII Congressus. Internationalis. Ornithologici. Academy of Sciences of the USRR, Moscow, i. 338–47. Bonifay, M.-F. and Brugal, J.-P. (1996), Biogéographie et biostratigraphie des grandes faunes du Pléistocène inférieur et moyen en Europe du Sud: apport des gisements français. Paléo 8: 19–30. Brodkorb, P. (1971), Origin and evolution of birds, in D. S. Farner and J. R. King (eds.), Avian Biology. Academic Press, New York, 19–55. Brook, B. W. and Bowman, D. M. J. S. (2004), The uncertainty blitzkrieg of Pleistocene megafauna. Journal of Biogeography 31: 517–23. Burney, D. A. and Flannery, T. F. (2005), Fifty millennia of catastrophic extinctions after human contact. Trends in Ecology and Evolution 20: 395–401. Catsadorakis, G. (2003), Greece’s heritage from nature. WWF, Athens. Chaline, J. (1972), Les rongeurs de l’Aven I des abîmes de la Fage à Noailles (Corrèze). Nouvelles Archives du Muséum d’Histoire Naturelle de Lyon, 10: 61–78. Chauvet, J.-M., Brunel Deschamps, E., and Hillaire, C. (1995), La Grotte Chauvet. Seuil, Paris. Cheylan, G. (1991), Patterns of Pleistocene turnover, current distribution and speciation among Mediterranean mammals, in R. H. Groves and F. di Castri (eds.), Biogeography of Mediterranean Invasions. Cambridge University Press, Cambridge, 227–62. Cooper, S. J. B., Ibrahim, K. M., and Hewitt, G. M. (1995), Postglacial expansion and genome subdivision in the European grasshopper Chorthippus parallelus. Molecular Ecology, 4: 49–60. Cowling, R. M., Rundel, P. W., Lamont, B. B., Arroyo, M. K., and Arianoutsou, M. (1996), Plant diversity in Mediterraneanclimate regions. Trends in Ecology and Evolution 11: 362–6. Crivelli, A. J. and Maitland, P. S. (1995), Endemic Freshwater fishes of the Northern Mediterranean Region. Biological Conservation 72: 1–337. Davis, M. B. (1976), Pleistocene biogeography of temperate deciduous forests. Geosciences Man 13: 13–26.

161

Davis, S. J. M. (2003), The zooarchaeology of Khirokitia (Neolithic Cyprus), including a view from the mainland, in J. Guilaine and A. Le Brun (eds.), Actes du Colloque International, Nicosie, 17–19 Mai 2001, 258–68. Dercourt, J., Zonenshain, L. P., Ricou, L.-E., Kazmin, V. G., Le Pichon, X., Knipper, A. L., Grandjacquet, C., Sbortshikov, I. M., Geyssant, J., Lepvrier, C., Pechersky, D. H., Boulin, J., Sibuet, J.-C., Savostin, L. A., Sorokhtin, O., Westphal, M., Bazhenov, M. L., Lauer, J.-P., and Biju-Duval, B. (1986), Geological evolution of the Tethys belt from the Atlantic to the Pamir since the Lias. Tectonophysics, 123: 241–315. Diamond, J. M. (1992), Twilight of the pygmy hippos. Nature 359: 15. (2002), Evolution, consequences, and future of plant and animal domestication. Nature 418: 700–7. di Castri, F. (1981), Mediterranean-type shrublands of the world, in F. di Castri, D. W. Goodall, and R. L. Specht (eds.), Mediterranean-Type Shrublands. Ecosystems of the World 11. Elsevier, Amsterdam, 1–52. Drake, J. A., Mooney, H. A., di Castri, F., Groves, R. H., Kruger, F. J., Rejmanek, M., and Williamson, M. (1989), Biological Invasions: A Global Perspective. John Wiley & Sons, Chichester. Dupouey, J. L., Dambrine, E., Moares, C., and Lafitte, J. D. (2002), Irreversible impact of past land use on forest biodiversity. Ecology 83: 2978–84. Elorza, M. and Sanchez Marco, A. (1993), Postglacial fossil Great Auk and associated avian fauna from the Biscay Bay. Munibe (Antropologia-Arkeologia) 45: 179–85. Ferrer, M. and Donazar, J. A. (1996), Density-dependent fecundity by habitat heterogeneity in an increasing population of Spanish Imperial Eagles. Ecology 77: 69–74. Georgoudis, A. (1995), Animal genetic diversity plays important role in Mediterranean agriculture. Diversity 11: 16–19. Gomez-Campo, C. (1985), Plant Conservation in the Mediterranean Area. Dr W. Junk, Dordrecht. Gregory, R. D., Greenwood, J. J. D., and Hagemeijer, E. J. M. (1998), The EBCC Atlas of European Breeding Birds: a contribution to science and conservation. Biologio e Conservazione della Fauna 102: 38–49. Grosbois, V., Henry, P., Blondel, J., Perret, P., Lebreton, J., Thomas, D. W., and Lambrechts, M. M. (2006), Climate impacts on Mediterranean blue tit survival: an investigation across seasons and spatial scales. Global Change Biology 12: 2235–49. Hagemeijer, W. J. M. and Blair, M. J. (1997), The EBCC Atlas of European Breeding Birds. T. and A. D. Poyser, London. Hewitt, G. M. (1996), Some genetic consequences of ice ages, and their role in divergence and speciation. Biological Journal of the Linnean Society 58: 247– 76. (1999), Post-glacial recolonisation of European biota. Biological Journal of the Linnean Society 68: 87–112. (2000), The genetic legacy of Quaternary ice ages. Nature 405: 907–13. Hino, T. (1990), Palaearctic deciduous forests and their bird communities: Comparisons between East Asia and WestCentral Europe, in A. Keast (ed.), Biogeography and Ecology of Forest Bird Communities. SPB Academic, The Hague, 87–94. Huntley, B. (1988), European post-glacial vegetation history: a new perspective, in H. Ouellet (ed.), Acta XIX Congressus Internationalis Ornithologici. National Museum of Natural Sciences, Ottawa, i. 1061–77.

162

Jacques Blondel

Huntley, B. and Birks, H. J. B. (1983), An Atlas of Past and Present Pollen Maps for Europe: 0–13000 Years Ago. Cambridge University Press, Cambridge. and Webb III, T. (1989), Migration: species’ response to climatic variations caused by changes in the earth’s orbit. Journal of Biogeography 16: 5–19. Jaeger, J. J. (1975), The mammalian faunas and hominid fossils of the Middle Pleistocene of the Maghreb, in K. W. Butzer and G. L. Isaac (eds.), After the Australopithecines. The Hague, Mouton, 68–75. Jahn, G. (1991), Temperate deciduous forests of Europe, in E. Röhrig and B. Ulrich (eds.), Temperate Deciduous Forests. Ecosystems of the World 7. Elsevier, Amsterdam, 377–502. Klicka, J. and Zink, R. M. (1997), The importance of recent Ice Ages in Speciation: a failed paradigm, Science, 277: 1666–9. Kowalski, K. and Rzebik-Kowalska, B. (1991), Mammals of Algeria. Polish Academy of Sciences, Wroclaw. Kurtén, B. (1968), Pleistocene Mammals of Europe. Weidenfeld & Nicolson, London. Kvist, L., Viiri, K., Dias, P. C., Rytkönen, S., and Orell, M. (2004) Glacial history and colonization of Europe by the Blue tit (Parus caeruleus). Journal of Avian Biology 35: 352–9. Latham, R. E. and Ricklefs, R. E. (1993), Continental comparisons of temperate-zone tree species diversity, in R. E. Ricklefs and D. Schluter (eds.), Species Diversity in Ecological Communities: Historical and Geographical Perspectives. Chicago University Press, Chicago, 294–314. Le Houérou, H.-N. (1990), Global change: vegetation, ecosystems, and land use in the southern Mediterranean Basin by the mid twenty-first century. Israel Journal of Botany, 39: 481–508. Lister, A. M. (1996), Dwarfing in island elephants and deer: processes in relation to time and isolation. Symposium of the Zoological. Society of London 69: 277–92. McGlone, M. S. (1996), When history matters: scale, time, climate and tree diversity. Global Ecology and Biogeography Letters 5: 309–14. McNeil, J. R. (1992), The Mountains of the Mediterranean World, an Environmental History. Cambridge University Press, Cambridge. Martin, P. S. (1984), Prehistoric overkill: the global model, in P. S. Martin and R. G. Klein (eds.), Quaternary Extinctions. University of Arizona Press, Tucson, 354–403. Michaux, J. R., Libois, R., Ramalhinho, M. G., and Maurois, C. (1998), On the restriction patterns and variation of the Iberian wood mouse (Apodemus sylvaticus). Comparison with other west Mediterranean populations. Hereditas 129: 187–94. Mönkkönen, M. (1994), Diversity patterns in Palaearctic and Nearctic forest bird assemblages. Journal of Biogeography 21: 183–95. Moritz, C. (1994), Applications of mitochondrial DNA analysis in conservation: a critical review. Molecular Ecology 3: 401–11. and Faith, D. P. (1998), Comparative phylogeography and the identificaion of genetically divergent areas for conservation. Molecular Ecology 7: 419–29. Mourer-Chauviré, C. (1975), Les Oiseaux du Pléistocène Moyen et Supérieur de France. Documents du Laboratoire de la Faculté des Sciences de Lyon, 64: 1–624. and Antunes, M. T. (1991), Présence du Grand Pingouin, Pinguinus impennis (Aves, Charadriiformes) dans le Pléistocène du Portugal. Geobios 24: 201–5.

Oosterbroek, P., and Arntzen, J. W. (1992), Area-cladograms of Circum-Mediterranean taxa in relation to Mediterranean palaeogeography. Journal of Biogeography 19: 3–20. Owen-Smith, N. (1987), Pleistocene extinctions: the pivotal role of megaherbivores. Paleobiology 13: 351–62. Pierpaoli, M., Riga, F., Trocchi, V., and Randi, E. (1999), Species distinction and evolutionary relationships of the Italian hare (Lepus corsicanus) as described by mitochondrial DNA sequencing. Molecular Ecology 8: 1805–17. Pons, A. (1981), The history of the Mediterranean shrublands, in F. Di Castri, D. W. Goodall, and R. L. Specht (eds.), Mediterranean-Type Shrublands. Ecosystems of the World 11. Elsevier, Amsterdam, 131–8. Quézel, P. (1985), Definition of the Mediterranean region and origin of its flora, in C. Gomez-Campo (ed.), Plant Conservation in the Mediterranean Area. Dr. W. Junk, Dordrecht, 9–24. Rage, J. C. (1972), Les amphibiens et les reptiles du gisement des abîmes de la Fage. Nouvelles Archives du Muséum d’Histoire Naturelle de Lyon 13: 113–17. Reyjol, Y., Hugueny, B., Pont, D., Bianco, P. G., Beier, U., Caiola, N., Casals, F., Cowx, I., Economou, A., Ferreira, T., Haidvogl, G., Noble, R., De Sostoa, A., Vigneron, T., and Virbickas, T. (2006), Patterns in species richness and endemism of European freshwater fish. Global Ecology and Biogeography 16: 65–75. Rögl, F. and Steininger, F. F. (1983), Von Zerfall der Tethys zu Mediterran und Paratethys. Die neogene Palaeogeographie und Palinspastik des zirkum-mediterranen Raumes. Annals Naturhistorisch Museum Wien (A) 85: 135–63. Ronquist, F. (2002), Parsimony analysis of coevolving species associations, in R. D. M. Page (ed.), Cospeciation. Chicago University Press, Chicago, 22–64. Sanmartin, I. (2003), Dispersal vs. vicariance in the Mediterranean: historical biogeography of the Palearctic Pachydeminae (Coleoptera, Scarabaeoidea). Journal of Biogeography 30: 1883–97. Enghoff, H., and Ronquist, F. (2001), Patterns of animal dispersal, vicariance and diversification in the Holarctic. Biological Journal of the Linnean Society 73: 345–90. Seligman, N. G. and Perevolotsky, A. (1994), Has intensive grazing by domestic livestock degraded Mediterranean Basin rangelands?’ in M. Arianoutsou and R. H. Groves (eds.), Plant– Animal Interactions in Mediterranean-Type Ecosystems. Dr. W. Junk, Dordrecht, 93–104. Simmons, A. H. (1988), Extinct pygmy hippopotamus and early man in Cyprus. Nature 333: 554–7. Sket, B. (1999), The nature of biodiversity in hypogean waters and how it is endangered. Biodiversity and Conservation 8: 1319–38. Steinfartz, S., Veith, M., and Tautz, D. (2000), Mitochondrial sequence anlaysis of Salamandra taxa suggests old splits of major lineages and postglacial recolonizations of Central Europe from distinct source populations of Salamandra salamandra. Molecular Ecology 9: 27–36. Suc, J. P. (1984), Origin and evolution of the Mediterranean vegetation and climate in Europe. Nature 307: 429–32. Taberlet, P. and Bouvet, J. (1994), Mitochondrial DNA polymorphism, phylogeography, and conservation genetics of the

Nature and Origin of Vertebrate Fauna brown bear Ursus arctos in Europe. Proceedings of the Royal Society, London B 255: 195–200. Fumagali, L., Wust-Saucy, A.-G., and Cosson, J.-F. (1998), Comparative phylogeography and postglacial colonization routes in Europe. Molecular Ecology 6: 289–301. Tchernov, E. (1984), Faunal turnover and extinction rate in the Levant, in P. S. Martin and R. G. Klein (eds.), Quaternary Extinctions. University of Arizona Press, Tucson, 528–52.

163

(1992), Eurasian-African biotic exchange through the Levantine corridor during the Neogene and Quaternary. Courier Forsch.-Inst. Senckenberg 153: 103–23. Voelker, G. (1999), Dispersal, vicariance, and clocks: historical biogeography and speciation in a cosmopolitan passerine genus (Anthus: motacillidae). Evolution 53: 1536–52. Webb, T. and Bartlein, P. J. (1992), Global changes during the last 3 million years: climatic controls and biotic responses. Annual Review of Ecology and Systematics 23: 141–73.

This chapter should be cited as follows Blondel, J. (2009), The nature and origin of the vertebrate fauna, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 139–163.

This page intentionally left blank

II

Process and Change in Specific Environments

This page intentionally left blank

Editorial Introduction Jamie Woodward

The nine chapters in Part II build on the physical, biological, and theoretical frameworks set out in Part I, but with a focus on process regimes and change in specific environments. With its emphasis on much larger spatial scales, Part I showed how the Mediterranean basin is a product of long-term interactions between all components of the Earth system. It showed how these interactions drive landscape and ecosystem processes and environmental change. The chapters in Part II examine Mediterranean-wide patterns too, but explore process interactions in sharper resolution and across scales ranging from individual soil profiles, hillslopes, and habitats to larger landscape elements including lake basins, river valleys, dune systems, and coastal plains. Much of the region is dominated by mountains and many process interactions are especially vivid in the Mediterranean because of the erosive energy available in steep and active tectonic settings, and the presence of soft rocks vulnerable to mass movements and water erosion. Abrupt transitions from uplands to lowlands— and the differential response to tectonic uplift of hard and soft rock terrains—are notable features. The seasonally dry climate can leave bare slopes exposed to high intensity rains, and river sediment yields are typically much higher than in adjacent regions. It can be argued that the Quaternary records of these interactions are more varied and better preserved than in any other part of the world. Recent major advances include the development of high resolution proxy climate data from speleothems and robust dating frameworks for fluvial, glacial, and palaeoecological records. These records have provided important new insights into the tempo of climate, landscape, and ecosystem change in the Mediterranean region and beyond. A variety of sedimentary archives also provide insights into the changing nature and intensity of

human action in the Mediterranean landscape during the course of the Pleistocene and Holocene and this is a core theme of Part II. The region is unique because of the very early and widespread impact of humans in landscape and ecosystem change—and the richness of the archaeological and geological archives in which it is chronicled. As several chapters show, however, disentangling the human impact from other drivers of change is not straightforward. A good deal of the research on Quaternary landscape change has been conducted within archaeological projects and key outcomes include the emergence of landscape archaeology and a developing tradition of geoarchaeological research across the Mediterranean; from the Palaeolithic through to Classical and more recent times. The chapters of Part II present a region of great diversity and stark contrasts. Together they show that searching for basin-wide, unifying characteristics is, ultimately, a futile exercise. But this does not diminish the importance of considering a Mediterranean-wide physical geography. This need is well illustrated throughout Part II by the myriad of material transfers that take place within and beyond the various subsystems of the basin. Such transfers (and transformations) are key elements of this physical geography and include, for example, fluvial sediment transfer from the mountains to the coastal zone; karst solution and precipitation; the flux of aeolian dust from the Sahara to the sea and to the soil profiles of southern Europe. What emerges from these chapters is a dynamic and varied physical geography— with a network of material and energy fluxes operating within and between markedly contrasting settings. A better understanding of such fluxes, their underlying controls (including the human impact), and their variation over time will lead to a better understanding of the Mediterranean environment, past, present, and future.

This page intentionally left blank

6

Weathering, Soils, and Slope Processes John Wainwright

Introduction Hillslopes are the dominant landform features of the Earth’s surface. They make up the interface between the atmosphere and Earth systems, providing a substrate that supports life and thus the basis for human activities within the Mediterranean. Their location at this interface means that hillslopes evolve through a complex interaction of different processes, operating at a range of different time and spatial scales. At longer timescales, processes of weathering convert rock and other parent materials into soils. Soils allow the growth of vegetation and thus further feedbacks between atmospheric and surface processes; in some cases these feedbacks can be seen to provide relative stability, while in others the system can become more fragile (Chapter 20). The latter case often arises as a result of erosion processes of various types. Water erosion and mass movements are a significant element of Mediterranean landscape evolution, occurring in parallel with (in response to, and affecting) tectonic processes that have moulded the configuration of the Earth’s crust (see Chapter 1), producing the unique combination of environmental characteristics of the region. Since the Late Pleistocene, depending on location, human activity has led to an acceleration of many of these processes, with important consequences for the basic ‘life-support system’ of the region and for global environmental cycles.

Weathering Processes The in situ modification of near-surface materials is typically considered to take place along a continuum relating to the dominance of mechanical or chemical processes (e.g. Birkeland 1999). The simplest control may be considered to be climatic, with mechanical

breakdown of particles dominating in cold, dry conditions, and chemical processes dominating in warm, wet conditions. Comparing this model to the present day climate of the Mediterranean (Figure 6.1) suggests, as with other processes, something of a north–south divide in terms of the dominant weathering process. The northern part of the basin (together with the Levant and the north-facing uplands of the Maghreb) would seem to be dominated by moderate chemical weathering; exceptions being the arid areas of south-east Spain, southern Sicily, eastern Cyprus, and parts of the Anatolian plateau as well as areas where low average temperatures would also reduce rates, such as in the Alps and parts of Slovenia and Croatia. Strong mechanical weathering is only estimated as occurring in the eastern Alps and in eastern Turkey while strong chemical weathering is presently limited to the wetter areas of Galicia, Slovenia, and Croatia. The southern part of the basin—with the exception already noted and a small area in northern Libya—are estimated as undergoing only very slight weathering. The limiting factor in this case is one of aridity (see Chapter 3). The spatial pattern suggested by the climatic model suggests that, on the whole, modern rates of weathering should be moderate to slow, which would further suggest the presence of relatively shallow soils. While in many areas of the Mediterranean, this pattern is indeed observed, it hides a considerable amount of variability in the nature and type of soils. Pope et al. (1995) have argued that the climatic approach greatly oversimplifies the understanding of weathering, and that there is a need to concentrate on the local— even microscopic—scale to understand the types and patterns of weathering that occur. Even if temperature and precipitation are important controls on the process, the climatic conditions give only a broad representation

170

John Wainwright Mean annual precipitation (mm) 2000

1500

1000

500

(a)

Mean annual temperature (°C)

-12°

0° Mechanical weathering Strong Moderate Weak

12°

Chemical weathering Strong Moderate 24°

Very slight weathering Water

(b)

0

500 km

Fig. 6.1. Potential climatic controls on weathering processes: (a) relationship between annual average precipitation and temperature and strength of mechanical or chemical weathering as defined by Peltier (1950, as modified by Pope et al. 1995); and (b) spatial distribution of these dominant types as estimated for the period 1961–90 based on the 10’ gridded climate data of New et al. (2000). Note the discussion in the text about how well this distribution represents the rates at a specific location, especially in terms of altitudinal and other local variations. Modified from the cited sources.

Weathering, Soils, and Slope Processes

of what might occur; the weathering process itself takes place at microscales relating to mineral structures and grain boundaries, and on slightly larger scales along fractures. Thus it is necessary to focus on local scale characteristics such as the chemistry, composition, texture, and structure of parent materials, microclimate, topography, and microtopography, as well as the presence of abiotic and biotic weathering agents and feedbacks between these characteristics and the weathering process. Furthermore, Pope et al. (1995) emphasize the importance of the interaction between mechanical and chemical weathering, for example solution leading to the development of microfissures that can then be further opened by mechanical action (thus potentially providing a greater surface area for solution to occur). Given the highly variable nature of Mediterranean environments in terms of their geology, hydrology, microclimatology, and vegetation inter alia, it should be expected that the spatial pattern and type of weathering process is more highly variable than the pattern shown in Figure 6.1. The geological evolution of the Mediterranean basin (Chapter 1) has been dynamic and long lasting (see also the summary in Wainwright and Thornes 2003). While the opening of the Tethys Ocean produced large quantities of carbonate rocks, it also resulted in deeper water clastic sediments, such as flysch. Evaporitic rocks— especially relating to the Messinian Salinity Crisis (see Chapters 1 and 2) are also common. Tectonic processes have juxtaposed a large number of different terranes, as well as producing igneous rocks (basalts and other basic and ultrabasic rocks around spreading ridges as well as more intermediate and acidic volcanic rocks) and metamorphosing existing rocks. These dynamics have been part of (geologically) rapidly changing stress fields within the crust (Dercourt et al. 1986; Dewey et al. 1989), producing complex patterns of fracturing of rocks over large areas. Denudation rates of Mediterranean limestones vary over a relatively small range. Frumkin and Fischhendler (2005) suggest rates of about 10–30 mm ka−1 over the Plio-Pleistocene in central Israel in areas that currently experience 100–700 mm of annual rainfall. In northern Italy, rates of 20 mm ka−1 have been measured under 1,442 mm annual precipitation compared to 30 mm ka−1 under 2,800 mm (Forti 1984, cited in Kaufmann and Braun 2001). Serrat (1999) estimated a current denudation rate of 35.4 mm ka−1 in a catchment dominated by carbonates in the Pyrenean foothills of south-west France, where mean precipitation is 700 mm. These rates can be compared to measurements of 5–17 mm ka−1 made on bare rock surfaces (Jennings 1985), which support the idea of a

171

positive feedback with shallow soil depths because of the increased contact time with water and development of humic acids (Ahnert 1976). These denudation rates should be considered to be areal averages, including the spatial variability produced by preferential dissolution along fractures and the development of karst forms (e.g. Dreybrodt 1988; cf. Chapter 10). Mechanical weathering of carbonates can also be seen to be widespread especially in the mountains of Spain, southern France, Italy, and northern Greece, for example. García-Ruiz et al. (2001) carried out a detailed study of screes in the Spanish Pyrenees at altitudes of 700–1,000 m where present-day mean annual temperatures are about 12◦ C and precipitation from 800–1,250 mm. The main bodies of these screes are about 4–5 m thick, and seem to date to the Late Glacial period, although thinner deposits of about 0.5 m suggest slower mechanical weathering has been ongoing in the later Holocene. Similar deposits can be seen throughout the Mediterranean mountains, and deep gorges lined with thick scree deposits are common (Figure 6.2; Chapter 12). Dissolution producing karst is also typical of evaporitic bedrock, notably in Spain and Turkey. Do˘gan and Özel (2005) suggest dissolution of gypsum is on average 10 to 30 times more rapid than that of limestone, which is supported by their observations in Central Anatolia. In the Sorbas basin of south-east Spain, there has been significant recent dissolution of the surface gypsum, despite the relative aridity of the area (300– 400 mm) as well as ongoing karstification through the Pleistocene (Calaforra and Pulido-Bosch 2003). Abandoned quarry sites in this same location were used by Dana and Mota (2006) to demonstrate the development of 50–60 mm of soil over a 10–12 year period compared to 70–80 mm over a 70-year period, again demonstrating non-linearities in process rates through time. The weathering of igneous bedrocks tends to be more rapid than that of the carbonates and evaporites, largely due to the greater chemical disequilibrium of the parent material, but is also in part structurally controlled, due to the formation of fissures during cooling, for example. Scarciglia et al. (2005) describe poorly differentiated weathering profiles from one to tens of metres deep in granites in the mountains of Calabria. Detailed analysis of the microscopic and macroscopic forms of weathering, as well as chemical analyses demonstrates the interplay of mechanical and chemical weathering at this location, where the mean annual rainfall is 1,400–1,600 mm, and the mean January temperature is −0.7◦ C. Dissolution of granites in the central Pyrenees may be as high as 8 mm ka−1 , although this value seems to be one of the highest reported for such systems (Oliva et al. 2004).

172

John Wainwright

Fig. 6.2. Gorges of the River Hérault in south-west France. Scree slopes can be seen on the long slopes to the left of the image and as aprons below the cliffs in the central part of the photograph (photo: John Wainwright).

Sanjurjo et al. (2003) demonstrate the importance of the interaction with vegetation in the development of soils on Mount Etna. Plantations in the 1940s and 1960s were able to produce soil profiles of over a metre in thickness, in areas where no soil had previously existed following the eruptions of volcanic cones in AD1669 and c.2,000 BP. In contrast, metamorphic denudation

rates tend to be relatively slow, producing clay-rich mineral assemblages. Van Wesemael et al. (1995) investigated soils on phyllites in Tuscany, where in situ soil profiles are typically of only a few tens of centimetres, and chemical weathering is reduced, probably due to low organic matter production under Mediterranean forests.

Weathering, Soils, and Slope Processes

A final broad grouping of geological control can be considered to be the poorly consolidated sediments— either marls or recent deposits that have not undergone significant lithification due to recent uplift. Marls such as the terres noires that are extensively found in southeastern France weather rapidly, especially in locations where bedding is at an angle to the ground surface (Bufalo et al. 1989; Malet et al. 2003). These soils and the clays of eastern and central Italy tend to be fine-grained and are subject to shrink–swell activity. Unconsolidated alluvial sediments on the other hand tend to be much coarser. Both contexts can produce deep soils in the case where erosion has not subsequently removed materials. The significance of vegetation has also been demonstrated in two cases. Vegetation can have both positive and negative mechanical effects, the former in terms of stabilizing the surface (e.g. van Beek et al. 2005), while the latter in terms of mechanical erosion from roots prising open fissures within individual particles or within the bedrock. Root systems in many Mediterranean plants can be extensive, especially in order to obtain moisture from depth during the summer drought (Chapter 7). Thus, a positive feedback can form with vegetation producing larger fissures which contain more water and more organic matter, leading to an acceleration of the weathering rate in the fissure, and its subsequent enlargement. Garty (1992) has demonstrated the importance of chelation on surface rocks in Israel recolonized after wildfires. In this study, it was estimated that unicellular green algae could remove 3.06 kg ha−1 of stone and that endolithic lichens could remove 174 kg ha−1 in the first few years of colonization. This process of colonization is important in the initial breakdown of bare rock surfaces, producing soil material that can then be colonized by propagules from higher-order plants. van Wesemael (1993) demonstrated that rates of litter formation in Mediterranean forest (either deciduous or sclerophyllous) was comparable to that under temperate forest, but that under Mediterranean conifers (Pinus pinaster), the rate was approximately 2.5 times slower, with comparable differences in humus development (and humic acid, so affecting relative chemical weathering rates) (van Wesemael and Veer 1992). Rates of litter decay under Cistus and Myrtus maquis in Spain have been found to be similar to those of the deciduous or sclerophyllous forest (Fioretto et al. 2003), as has the decay of Arbutus unedo and Quercus coccifera maquis in Greece (Arianoutsou 1993). The generally slow production of humus in Mediterranean systems is thus the result of low litter production, related to low biomass, rather than significant differences in decay rates when compared to more temperate environments.

173

Fire is a significant process in Mediterranean environments (Naveh 1975; Wainwright and Thornes 2003; Chapters 7 and 19), with direct and indirect effects on weathering and soil-forming processes. Important processes are the fracturing of large boulders or spalling of smaller fragments due to stresses formed during heating or subsequent cooling (Shakesby and Doerr 2006). Following the Montagne de Sainte-Victoire fire in southeast France in 1989, Ballais and Bosc (1994) estimated that areas which undergo frequent fires in uplands could experience between 10 and 100 times the weathering rate observed as being due to frost action, although they noted that sedimentary rocks were preferentially affected in this case. Indirectly, burning can lead to a significant reduction of the organic matter content of soils (Giovannini et al. 1987). Other effects are strongly temperature-dependent, so that Giovannini et al. (1988) observed little change in sand content for temperatures up to 170◦ C, but between 220 and 460◦ C there was a rapid rise so that sand became the dominant particle size due to aggregation and fusing of particles. Little further changes were observed at higher temperatures. Feedbacks via soil nutrients and plant growth are more complex. Total soil N and organic P drop rapidly above temperatures of 170◦ C, although there is a slight increase in inorganic P (Giovannini et al. 1990). Plant growth including root biomass and extension, however, seem to be encouraged by moderate intensity burning of the soil (between 220 and 460◦ C), but are reduced at higher intensities. Intensities of burning can vary over very small spatial scales (Imeson et al. 1992; Lavee et al. 1995), so that a complex mosaic of post-fire weathering conditions can arise. Topography plays a role in controlling weathering processes at a range of scales. At a point, microtopography is important in controlling the specific pathways of water (e.g. Lavee et al. 1995). At the hillslope scale, topography affects catenary processes and the lateral linkage of materials by erosion (see below). Furthermore, the convexity of the hillslope affects water concentration, even if subsurface flows are not significant in many Mediterranean environments, especially if reinforced by spatial patchiness of other factors such as vegetation (Cameraat 2002), bedrock structure (Dreybrodt 1988), or microclimate (Bolle 1995). At larger scales, perhaps the most important control is altitude, especially given the extensive nature of the mountain ranges surrounding the Mediterreanean. The effect of physical weathering due to freeze–thaw action at altitude has already been discussed (see also Gallart et al. 2002). Palacios et al. (2003) have also discussed the importance of snow at altitude in producing

174

John Wainwright

nivation features above 2,000 m in central Spain. They argue that the Mediterranean mountains are particularly sensitive to this process because of the contrasting conditions between winter and summer and the dryness of the summer promotes low vegetation cover that accelerates the process. A final aspect that has already been touched on in several places is that of time. As soils develop, feedbacks occur that may accelerate (e.g. by increasing moisture availability as soil depths start to increase) or decelerate (e.g. by creating anaerobic and colder conditions as depths continue to increase, or as voids fill with illuvial clays) the weathering of parent materials. One consequence of these changes in rate through time is that it is important to appreciate when considering denudation rates calculated using short data sets that extrapolation is almost certainly not linear. Modelling frameworks (e.g. Kaufmann and Braun 2001) are thus likely to be the best way of integrating information about denudation at different timescales, when combined with studies that try to evaluate the appropriate form of change in rate through time. Another important aspect of change through time is that of varying external controls, notably that of climate—with the consequent changes in vegetation and other related feedbacks. The pattern of weathering processes according to climate as given above is short term (Figure 6.1 is based on data from 1961–90), whereas weathering processes typically should be considered over periods of thousands or even millions of years. A detailed discussion of climate change and variability is given in Wainwright and Thornes (2003), so the detail will not be discussed further here (see also Chapters 3, 4, and 9). In the period from at least 4.7 Ma to 3.2 Ma, temperatures were up to 3◦ C higher than at present. It seems likely that the western Mediterranean had more moisture than currently, and was thus under subtropical conditions. Higher rates of chemical weathering at this time are indicated by mineralogical analysis (Saez et al. 2003) and by the presence of deep, lapiez-style karst and pinnacles now exposed in southern Spain and France. In the central and eastern Mediterranean, higher temperatures seem to have been combined with stressed moisture regimes as at present, suggesting that there was a less marked difference in chemical weathering rate in these areas. However, Yassoglou et al. (1997) point out the presence of similar, deep soils prior to the Late Pleistocene in Greece, and Atalay (1997) who interprets red Mediterranean soils in southern Turkey as palaeosols, so at least locally, chemical weathering is also likely so have been more important in this part

of the basin. Schulte and Julia (2001) suggest that Early Pleistocene parts of soil chronosequences in the south of Spain are also significantly redder than later soils, again suggesting heightened amounts of chemical weathering. Progressive cooling continued, possibly with marked declines in temperature at around 2.5 Ma and 900 ka, with large amplitude oscillations set in place by the end of this period. While the total area affected by glaciation outside the Alps was relatively small (Hughes et al. 2006; see Chapter 12), the largescale cooling would mean that more frequent freezing conditions at lower altitudes and latitudes would have taken place. An important consideration based on analysis of the 975-ka pollen diagram from Tenaghi Philippon (Mommersteeg et al. 1995) is that the colder conditions were the most dominant—lasting for 400 ka out of the last 975 ka (Wainwright et al. 1999). One consequence was probably more widespread mechanical weathering, which in part may explain the widespread stony nature of Mediterranean soils (see also below) and thick accumulations of angular limestone clasts in caves and rockshelters at lower elevations (Woodward and Goldberg 2001; see also Chapter 10). Simon et al. (2000) demonstrated, for example, that soil-forming episodes as demonstrated by chronosequences in the Sierra Nevada of southern Spain were characterized by strong mechanical weathering features. By comparison, conditions approximating those of the present day seem to have occurred for only around 322 ka, and a transitional cooler, wetter climate occurred for 253 ka of the 975-ka record. Given that conditions relating to greater mechanical weathering are likely to have dominated over the last million years, caution is required when interpreting the nature of Mediterranean soil-forming processes. All too frequently, consideration is given of modern-day conditions, when these have been a relatively recent (and infrequent) part of the Mediterranean Quaternary environment. In summary, while the climate approach can give a possible broad scale interpretation of weathering processes, the conceptual model of Pope et al. (1995) is a useful one in terms of interpreting processes at a specific location. Heterogeneity in lithology, structure, vegetation, topography, and fire regime, as well as spatial and temporal feedbacks mean that a bottom-up approach to understanding weathering processes is most appropriate. Measured soil depths as a function of climate and lithological variability (and thus also the other factors mentioned) on the island of Lesvos are a good indication of this heterogeneity (Figure 6.3: Kosmas et al. 2000).

Weathering, Soils, and Slope Processes Semi-arid zone

175

Dry sub-humid zone

100 Lava Pyroclastics Schist-marble Vitrophyric lava Peridotite Ignimbrite Shale

Probability

80

60

40

20

0 0-15

16-30

31-50

51-100

>100 0-15 Soil depth (cm)

16-30

31-50

51-100

>100

Fig. 6.3. Comparison of soil depths as measured in different climate zones and on different lithologies on the island of Lesvos (modified from Kosmas et al. 2000).

Mediterranean Soils Distribution and Types of Soil Yaalon (1997) has provided an overview of Mediterranean soils, considering their classification and genesis. While earlier soil-classification systems had used terms such as red and brown Mediterranean soils, this use largely fell out of practice from the later 1960s. In the revised FAO classification (see below), these soils are now considered to be luvisols or cambisols. A major element of the debate was the common perception of the dominance of red soils, which Yaalon considers to be related to the oxidation of iron oxides (Figure 6.4) under dry, summer conditions. As such, red soils from this source would have to be relatively recent, or relate to past conditions of similar conditions (e.g. Woodward et al. 1994), based on our knowledge of climate conditions. Another mechanism of reddening would be heating during wildfires, which can also bring about similar transformations to haematite at temperatures between 200 and 400◦ C. Recent work by Wondafrash et al. (2005) has suggested that relatively permanent changes in colour of the surface horizons of Mediterranean soils can come about through burning. In this case, the persistence of red-coloured horizons at depth in the soil will depend on the relative dissolution and erosion rates and fire frequency. While red soils—commonly using the term terra rossa—are generally associated with limestone bedrock,

it is now appreciated that red soils also occur on other lithologies, including igneous, metamorphic, and conglomerates (Yassoglou et al. 1997). MacLeod (1980) pointed out the problem of assuming the reddening of limestone-derived soils could come from in situ alteration of non-carbonate residues, given that the latter are typically very small (<2%). Thus, observed soil thicknesses were difficult to explain by this mechanism alone. MacLeod suggested that aeolian input from deflation from the Sahara (Chapter 14) was an important source of iron oxides, and Jackson et al. (1982) also argued that these could produce reddening of soils. While some authors have disputed the necessary cause of reddening as being the input of Saharan dust (for example, Boero and Schwertmann 1989; Moresi and Mongelli 1988), it is widely recognized that Saharan dust is a major input into northern Mediterranean soils. Based on measurements in Crete, Pye (1992) measured dust-deposition rates equivalent to 0.77 to 7.7 mm ka−1 , while Nihlen et al. (1995) estimated rates of 6.6 to 21.4 mm ka−1 . Isotopic analyses of speleothems in Israel suggest dust inputs and sources have persisted in this area over at least the last 220 ka (Frumkin and Stein 2004), although Larrasoana et al. (2003) suggested that input was reduced during phases when the ITCZ extended northwards during phases of insolation maxima, based on a 3-Ma record from deep sea sediments. Again, care is required in extrapolating short-term measurements to longer-term landscape-formation processes. Inputs of Saharan dust to the central (Correggiari et al. 1989;

176

John Wainwright

Dissolved 3-

Fe

Hexagonal-close packing Cubic-close packing

Deprotonation

3-x

Fe (OH)2

Hydrolysis, Nucleation, Crystalization +Cl–

pH5.1

Akaganeite b-FeOOH

Ferrihydrite Fe3HO34H 2O

Dehydration Rearrangement

Haematite a-Fe2O2

Thermal transformation 200–400°C

pH5-7

Maghaemite g-Fe2O2

Thermal Trans.

Dehydroxylation

Oxidation Green rust s pH5- 7

Lepidocrocite g-FeOOH

Magnetite Fe3O4

Thermal transformation

Oxidation + carbonate

Goethite a-FeOOH

pH5.2

Feroxyhyte d-FeOOH

Oxidation + dehydration pH>8 3–x

Fe (OH)x Precipitation

3–g

Fe (OH)g

Rapid oxidation

Oxidation, Deprotonation

Fe2+

Fig. 6.4. Processes leading to the formation of iron oxides and thus the development of brown (goethite) and red (haematite) soils. Thicker arrows are the more common processes (modified from Yaalon 1997).

Tomadin et al. 1984, 1990) and north-western Mediterranean (Bücher 1989; Durand et al. 1992) have also been noted. Loess soils, where dust deposition exceeds about 25 mm ka−1 , can be found in Israel (Issar et al. 1989), Tunisia, and Morocco (Thomas 1997). However, measurements of present-day dust accumulation suggest that the loess is not actively forming, but again relates to past fluctuations in the position of the ITCZ. Figure 6.5 shows the spatial distribution of soils in the Mediterranean according to the FAO classification. While there is again a north–south distinction in the

types of soil present, the pattern (even as greatly simplified in this map) is more complicated than that of the climate-driven process map (Figure 6.1), providing further support for the Pope et al. (1995) model of local conditions controlling the type of soil-forming process. In the north, cambisols are the most frequent type (Table 6.1), dominated by eutric, calcic, and dystric cambisols in almost equal measure. Luvisols are next in importance, with chromic, orthic, and then gleyic luvisols being the main types present. Lithosols are the next most frequent type, followed by fluvisols (mainly calcaric) and xerosols (again mainly calcic). No other soil type covers more than 4 per cent of the total area. In the south, yermosols (dominantly calcic followed by undifferentiated with a small proportion of gypsic yermosols) are most frequent, followed by lithosols and podzoluvisols. Xerosols (mainly calcic) are around twice as frequent as (mainly calcic) fluvisols—mostly in the Nile valley—and cambisols. The latter are less than ten times as frequent in the south as in the north. These soil types make up nearly 90 per cent of the total area in the south. It should be clear that Yaalon’s (1997) description of ‘typical’ Mediterranean soils relates only to the northern part of the basin, and even so, the notion of a typical soil conceals a great deal of variability. Figure 6.6 illustrates a number of example soil profiles from the northern part of the basin. It should be no surprise that soils in the south are poorly developed, relating to the arid regime (almost 50% are yermosols or xerosols, which are characterized by a weakly developed ochric A horizon and a further 17% are lithosols, which are characterized by having a hard rock level within 10 cm of the surface) and the higher prevalence of wind erosion (Chapter 14). Comparison of the climate-based characterization and the FAO soil classification shows some convergence—for instance the majority of cambisols and luvisols occur in the moderate chemical weathering area, and most yermosols are in the very slight weathering zone—but also some differences. Nearly two-thirds of the lithosols are in the very slight weathering area, with just over a third in the moderate chemical weathering class; very few occur in zones characterized climatically as being dominated by mechanical weathering. While this pattern may in part be due to the scale of mapping and inadequacies with the climate model, it may also suggest that these soils are relict from past climate regimes. Alternatively, their location, which is dominantly in an arc through the Balkans and into Asia Minor and southern Turkey may reflect areas where high winter precipitation biases the climate model. Almost as many xerosols occur in areas mapped as undergoing moderate chemical

Weathering, Soils, and Slope Processes

177

Fig. 6.5. Spatial distribution of soils in the Mediterranean basin according to the FAO classification scheme (modified from Wainwright and Thornes 2003, based on data provided by NOAA 1984).

TABLE 6.1. Comparison of proportions of different soil types according to the FAO classification relative to location in the north or south of the Mediterranean basin North Cambisols Luvisols Lithosols Fluvisols Xerosols Rendzinas Chernozems Kastanozems Podzols Phaeozems Regosols Vertisols Rankers Planosols Andosols Gleysols Acrisols Solonchaks Podzoluvisols

%

South

%

42.1 16.4 8.9 4.6 4.3 3.5 3.5 3.1 2.4 2.3 2.1 2.0 1.0 0.8 0.7 0.6 0.6 0.3 0.3

Yermosols Lithosols Podzoluvisols Xerosols Fluvisols Cambisols Luvisols Solonchaks Regosols Rock Rendzinas Kastanozems Vertisols Salt

41.2 16.7 15.2 8.1 4.3 3.9 2.6 2.2 1.8 1.0 1.0 0.9 0.5 0.2

Source: Analysis is based on the NOAA-AVHRR (1984) data (after Wainwright and Thornes 2003).

weathering as being subject to very slight weathering. These possible anomalies are located in the northern Atlas and in central Anatolia. The idea that soil-forming processes have changed significantly through time has already been considered in terms of controlling mechanisms. Further lines of support for this idea come from the study of palaeosols. For example, Leone et al. (2000) have described a 150-m stratigraphic section in central Italy that is interpreted to have been deposited over a period of 100 to 300 ka in the Late Pliocene, which contains seventy-three palaeosols as well as lignite layers. Pollen evidence shows the presence of subtropical forest species, although isotopic analyses suggest that precipitation was not much higher than at present, and also points to the presence of grass vegetation. Günster and Skowronek (2001) have investigated sequences in the Granada basin of southwest Spain that contain sixty-five palaeosols within a thickness of up to 70 m of alluvial fan sediments. Early Pliocene and Early Pleistocene palaeosols show higher redness values, suggestive of more subtropical conditions, while Later Pliocene soils are more similar to those

178

John Wainwright

a Ah BT1 BT2 C/R

b

c

d

Ap

Ap

Ah

A

Bw

Apcle1

Dominant particle size: clay silt sand

g

Apcle2

B

0 cm 10 20 30

f

Apcle1

A B

e

pores Bk1

B1 B2

h Ap1 Ap2 Ah pores Bw Bwk1 Bwk2

AC

Bk2 Bk3

C CaCO3 concretions

ACck2 CaCO3 nodules

Bt3

CaCO3 accumulations

Bwk3 BkC

Ck

2C ‘mottles’

Fig. 6.6. Examples of typical soil profiles from the Mediterranean, as described by Tavernier (1995): (a) terra rossa on a cherty limestone near Rieti in Italy; (b) a pellic vertisol derived from igneous rocks in southern Portugal; (c) a ferric luvisol derived from schists in Portugal; (d) a ‘Black Mediterranean Soil’ (pellic vertisol) derived from marls near Florence in Italy; (e) a calcaro-pellic vertisol derived from Tertiary, calcareous lacustrine deposits in Thrace, Greece; (f) a calcaro-vertic cambisol derived from lagoonal clays in south-east Italy; (g) a calcivertic luvisol derived from Miocene clays near Badajoz, Spain; and (h) a calcic xerosol derived from Quaternary alluvium in Murcia, Spain (modified from Wainwright and Thornes 2003).

actively forming and contain up to sixteen pedogenic calcretes, which would also suggest at least seasonal aridity. In the Pleistocene, Ortiz et al. (2002) investigated soils on relatively stable alluvial fan surfaces in this same basin. They found that soils formed up to about 240 ka had higher amounts of clay and iron oxides, suggesting greater pedogenic activity during warmer periods. In some of the older soils, micromorphological evidence of frost shattering suggested mechanical weathering that the authors attributed to activity in cold periods (a feature also noted by Günster and Skowronek 2001, in mid–Late Pleistocene palaeosols in the area). Furthermore, older soils had greater CaCO3 accumulations in the C horizon, interpreted as being due to infiltration with carbonate-rich runoff, although input of carbonate from dust cannot be excluded. Indeed, Pulido-Villena et al. (2006) have measured average inputs of 1.6 g m−2 a−1 of Ca in high mountain lakes in Spain. Late Pleistocene–Early Holocene palaeosols in Israel (Gvirtzman and Wieder 2001) have been related to more humid phases relating to ITCZ fluctuations as discussed above. In the Pyrenean scree slopes already discussed, García-Ruiz et al. (2001) ascribe a phase of carbonate cementation to warmer, more humid conditions in the early Holocene. Van Andel and

Runnels (1987; van Andel et al. 1990) also describe early Holocene palaeosols from different areas in southern Greece.

Soils as a Resource As noted by Sumner and Wilding (2000: p. xvii): ‘We owe our existence to the extremely thin but precious skin, called soil, which covers the unweathered and partially weathered geological formations at the Earth’s surface with a unique and extremely thin, fragile layer.’ Soils store water and provide the physical and nutrient support for vegetation (Chapter 7) and thus animals (Chapter 5). As has been already noted, where older soils are not preserved, Mediterranean soils tend to be rather thin—often only a few tens of centimetres thick, and some slopes are bare, bedrock surfaces devoid of any soil cover. One consequence of thin soils is that they have a relatively low water-holding capacity, which will tend to promote runoff and erosion (see below), which provides a positive feedback to maintain relatively thin soils. This feedback operates not only by the physical removal of sediment, but also by the removal of nutrients, which, combined with at least seasonal aridity, maintains lower vegetation cover that also tends to promote erosion (see discussion in Thornes 1985, 1990). However, as

Weathering, Soils, and Slope Processes

also discussed, Mediterranean soils are characterized by variability and it would be incorrect to state that all soils are poor. Based on the FAO classification, 16 per cent of soils in the northern basin are classified as having a good nutrient status (mainly cambisols, but also some regosols and fluvisols), although not surprisingly, the value is only 1.3 per cent in the south. However, deep soils exist throughout the basin (approximately 4% of both the north and the south have fluvisols), in part as a response to deposition in areas of low relief of sediment eroded from uplands (where it is perhaps less useful as a resource), and a positive effect of the stoniness of soils is the maintenance of more humid soil microclimates (Casals et al. 2000), with stones acting as a mulch during periods of dryness (van Wesemael et al. 2000; Larcheveque et al. 2005). During the Late Pleistocene and through the Holocene, there has been increasing human impact on soils (Wainwright and Thornes 2003). The impacts of human activity on erosion will be discussed below, but it is important to recognize other negative and positive changes. Modification of the form of hillslopes through terracing and deliberate movement of soils (e.g. Gams et al. 1993; see also Figure 10.12 in Chapter 10) have been commonplace for millennia. These changes to the surface have the effect not only of increasing water-storage capacity and thus vegetation, but also tend to reduce the slope angle and thus minimize erosion (Van Andel and Runnels 1987). Traditional cultivation practices have often (but not always) maintained a more effective use of limited soils, and integrated agropastoral systems provide an effective means of recycling nutrients. Pollution of soils is also not restricted to the modern period. For example, Aberg et al. (2001) show evidence of soil contamination from lead smelting in the Hellenistic and Roman periods in Greece. It is clear that the pollution of soils has accelerated greatly since the mechanization of agriculture from the 1950s (e.g. Zalidis et al. 2002). Salinization is also a widespread (and long-lasting) problem, albeit often more accentuated in the southern and eastern parts of the basin. As elsewhere, pollution can also be related to distant sources. For example, Vavliakis et al. (1990) suggested that dissolution rates of dolomitic marbles in northern Greece doubled when dominant winds from eastern and central Europe increased the acidity of local rainfall to pH <5.1. Finally, not all pollution is anthropogenic in origin, with soils downwind of volcanoes that have significant amounts of degassing such as Etna having significant contents of fluorides (e.g. Davies and Notcutt 1988) and other toxic substances (Chapter 15).

179

Slope Processes Erosion of weathered or unconsolidated surface materials can be considered as a three-part process, involving the detachment of material from the surface or subsurface, transport, and deposition. The relative energy available to effect each of these subprocesses, together with the resistance afforded by the nature and organization of surface materials will control the rate at which erosion occurs, and thus how surfaces and ultimately landscapes change. It is useful to consider different mechanisms and their controls separately; in this case, consideration is given to diffusive processes, processes due to flowing water, and threshold processes producing instability.

Diffusive Erosion Processes Diffusive processes are typically the slowest within the landscape and result in gradual, generally step-wise movement of sediment away from a point, usually with a net downslope direction, although some processes can also result in some upslope movement. Soil creep occurs as a result of particle detachment by thermal expansion normal to the surface in the presence of matricial water and vertical transport by gravity (with deposition constrained by the location of adjacent particles), so that its rate is often related linearly to the slope angle. It is generally considered to be the slowest erosion process and to be ineffective in Mediterranean environments. However, Oehm and Hallet (2005) argue that local variability often overwhelms the dependence on slope angle, and that the climate control on creep rate also seems to be a weak one. Creep rates can be considered to be important in high mountain environments (where they merge into movements by solifluction or gelifluction, especially in colder periods of the Pleistocene). Even at lower altitudes, Wainwright (1996a) measured significant rates of creep in shallow soil profiles on steep limestone slopes (15–30◦ ) in south-western France (altitude 300 m a.s.l., annual precipitation 708 mm). Qualitative evidence such as displaced tree trunks can also often be seen on steeper slopes. Sorriso Valvo and Sylvester (1993) discuss the importance of creep where planes of weakness in weakly consolidated metamorphic rocks can be exploited in a mountainous area of Calabria. Creep is also an important precursor to other types of slope movement as demonstrated in the case of the Vaiont disaster discussed below. Erosion due to rainsplash occurs due to the direct impact of raindrops at the surface, which detaches particles that are either carried in rebounding splash droplets or directly through the air following ejection as the

180

John Wainwright

raindrop collapses laterally on impact. Particles follow a parabolic trajectory and are typically deposited within a few tens of centimetres from their origin, either upslope or downslope. There is typically a net downslope movement because of the geometry of the slope (Savat 1981), although Torri and Poesen (1992) have argued that microtopographic variations in the slope can be more important in controlling the direction of most transport. Splash can be important due to the high intensity (and thus energy) of rainfall, especially during convective storm events, and is particularly effective in detaching and transporting particles of fine sand size. This selectivity in transport can produce distinctive erosion pillars of several centimetres in height, capped by stones (Figure 6.7). The energy of raindrops reaching the surface is also affected by vegetation cover, which reduces the energy reaching the surface (unless tall canopies produce significant leaf drips, which is not usually the case given the structure of most Mediterranean tree species: Chapter 7). For example, Wainwright (1996b) noted a fivefold decrease in splash on grassed surfaces compared to bare surfaces in fields in south-eastern France, while Bochet et al. (2002) noted that rates under summerdeciduous shrubs were 51 per cent of bare soil rates, and the values for tussock grass and evergreen shrubs were

17 and 10 per cent, respectively, in Spanish matorral. One consequence of differential splash rates in the presence of relatively sparse vegetation is the development of mounds of fine sediments under shrubs, as the rate of splash into the area under a shrub canopy is less than the rate out of it (Parsons et al. 1992). Bochet et al. (2000) have observed this process in Spain, and it is analogous to the development of ‘islands of fertility’ as suggested for other dryland environments (Charley and West 1975), which produces feedback to the vegetation structure as the mounds provide higher water and nutrient contents available to the shrubs (Figure 6.8). Rates of erosion are also controlled by soil cohesion, which is inversely proportional to soil moisture, and other surface characteristics and their evolution through time. For example, Shakesby et al. (1993) found that recently burned soils under Eucalyptus and Pinus pinaster forest in northern Portugal had splash rates an order of magnitude higher than sites that were burned three to four years previously, and these were two orders of magnitude higher than sites that had not burned for ten years— although in part, vegetation-canopy recovery will also be an important control in this case. Once runoff occurs, splash rates drop rapidly (Torri et al. 1981; see examples of rates in Wainwright 1996b).

Fig. 6.7. Examples of splash pillars forming on marls (the ‘Terres Noires’) near Propiac in south-east France (see Wainwright 1996d, for more details of the erosion characteristics of this site). The lens cap in the foreground is 55 mm in diameter (photo: John Wainwright).

Weathering, Soils, and Slope Processes

181

Fig. 6.8. Examples of the development of patchy vegetation representing ‘islands of fertility’, from the Montpellier Garrigue, southern France; (a) under mixed bunch grass (Stipa sp.) and shrubs dominated by box (Buxus sempervirens and Cistus; and (b) detail under a box shrub. Note the pavement surfaces developed in the areas between vegetation clumps in both cases (photos: John Wainwright).

182

John Wainwright

A process that is increasingly recognized as being important in eroding soil is that of ploughing during agriculture. Tillage erosion occurs by detachment of the soil by agricultural equipment and its subsequent transport by gravity. Ploughing parallel to the contour will lead to both upslope and downslope movement— although the latter will be more important, and by analogy with splash, should relate to the slope angle. Slopeparallel ploughing will tend to produce predominantly downslope movement of soil, again in proportion to the slope angle (Figure 6.9). Depth of tillage (and thus essentially the extent to which the process is mechanized) is also an important control on the amount of sediment flux (De Alba et al. 2006). While some authors have argued that tillage erosion can be more important than erosion by overland flow (Govers et al. 1994), other authors have found the opposite pattern in Mediterranean environments (Hassouni and Bouhlassa 2006) and it is clearly important to consider local hydrological and surface conditions when evaluating the relative balance. There are also clear interactions between tillage regime and water erosion, as demonstrated, for example, by Gomez et al. (2004), with higher rates under conventional tillage than under light grass cover (Figure 6.10). Tillage can also increase surface roughness in the short term, which reduces the energy of water flows and can thus reduce water erosion rates (Cammeraat 2004). It also tends to increase the stoniness of the surface (Poesen et al. 1998), with consequences for the operation of other surface processes. One extreme where mechanical movement is undeniably higher than other forms of displacement is in the current practice of bulldozing terraces into slopes of uncon-

solidated sediment, as is commonly seen in southern Spain. Tillage erosion can be considered as a type of bioturbation of soils. Other processes of bioturbation may be less important in terms of absolute rates, but may also have a significant effect on slope properties and evolution. Bioturbation can in general be considered as a diffusive process, but the type of displacement will be a function of different behavioural patterns (burrowing, other types of digging, trampling, etc.). Bochet et al. (2000) have shown how burrowing activity can be an important part of soil displacement that leads to the development of mounds under Stipa and Rosmarinus shrubs. Yair (1995) compared soil movements by isopods and porcupines on slopes in Israel, and demonstrated that activity by both can lead to significant soil displacement, and that there is an important feedback with water erosion, in that the animals produce more easily detachable sediments for subsequent movement by water erosion. The different animals in Yair’s study had distinctive spatial patterns in their effects on erosion (Figure 6.11). Trampling by sheep has also been shown to be an important erosion process—especially of rock fragments—on slopes in Spain by Oostwoud Wijdenes et al. (2001). Their measurements estimated flux rates of 0.1 to 21 kg m−1 a−1 , depending on slope angle and vegetation cover.

Erosion Due to Flowing Water The characteristics of the Mediterranean region that tend to produce high amounts of runoff are discussed in Chapter 8. An important distinction in erosion from

(a)

(b)

100

120

Soil displacement dP (cm)

Soil displacement dT (cm)

120 y = -1.06x+13.50 R 2 = 0.48

80 60 40 20 0 -40 -35 -30 -25 -20 -15 -10 Slope gradient ST (%)

-5

0

5

100 80 60 40 20

y = -1.17x+52.73 R 2 = 0.40

0 -40 -35 -30 -25 -20

-15 -10

-5

0

5

Slope gradient SP (%)

Fig. 6.9. Measured soil displacements as a result of tillage erosion using a mouldboard plough in Tuscany: (a) slope-parallel tillage; and (b) contour-parallel tillage (modified from De Alba et al. 2006).

Weathering, Soils, and Slope Processes

183

Fig. 6.10. Accelerated erosion related to slope-parallel tillage in vineyards in southern France, following the storms of 12–13 November 1999 (photo: John Wainwright).

surface flows is that between unconcentrated and concentrated flows. Unconcentrated (or interrill) overlandflow erosion is often misleadingly called ‘sheet’ or ‘wash’ erosion. Observation of flows in the field demonstrates that the idea of a uniform sheet of flow does not occur due to microtopographic variations and the presence of obstacles such as vegetation and stones, as well as due to feedbacks within the flow structure itself. In unconcentrated overland-flow erosion, detachment of particles is predominantly by raindrop impact, although once flows enter the transitional regime, a limited amount of flow detachment occurs (see discussion in Wainwright et al. 2008). As noted above, raindrop detachment declines rapidly once flow starts to accumulate, so that a negative feedback occurs with the overland-flow erosion rate, irrespective of whether there is still erodible sediment at the surface. Transport distances tend to be of the order of tens of centimetres to a metre, depending on the flow energy and rainfall energy. Interrill erosion is a complex function of stone cover, which has been seen to be an important characteristic of many Mediterranean soils. Poesen (1994) demonstrated experimentally that

stones that sit on the surface will tend to promote infiltration and reduce erosion rates by up to an order of magnitude, whereas stones that are compressed into the surface promote runoff and can lead to up to a fivefold increase in interrill erosion (Figure 6.12). Crusting of the surface can affect erosion rates significantly especially on marls (Martínez Mena et al. 2002), where it can cause runoff to develop rapidly and flow at higher velocities over relatively smooth surfaces. Vegetation cover is again an important control on minimizing the energy for detachment and thus interrill erosion rates; typically, as canopy cover increases, erosion rates will decrease exponentially. Francis and Thornes (1990) demonstrated approximately an order of magnitude difference between erosion on sparse versus good matorral cover on slopes in southern Spain, and a further order of magnitude reduction under Pinus halepensis forest. Increasingly, as with other dryland environments, it is recognized that the spatial pattern of sparse vegetation canopies can be as important as cover itself in controlling erosion rates (Puigdefabregas et al. 1999; Kirkby et al. 2002; de Vente 2006: Figure 6.13). Factors such as

184

John Wainwright

(a) Porcupines A

B C D E F

G H I

J

K

L M

N

O

P Q R S T U V

(b) Porcupines & Isopods

W

66

44

68

36

50

Ne

37

tse

r

52 51

0

54

Sh

66

ivt

a

71

73

N

(a) 0 - 999

71

30 metres

66

Dr

56

68

or

(b)

im

60

67

53

54

64

58 41

36

0 - 1499

1000 - 1999

1500 - 2999

2000 - 2999

3000 - 4499

3000 - 3999

4500 - 5999

4000 - 4999

6000 - 7499

gm-2

% of easily erodible soil produced by porcupines per row

Fig. 6.11. Soil production by isopods and porcupines on hillslopes at Sede Boquer, southern Israel (modified from Yair 1995).

fire, grazing, or agriculture that can significantly affect vegetation cover therefore have a significant control on erosion rates. In the case of fire, there are often further feedbacks through the development of hydrophobicity (Shakesby and Doerr 2006), while grazing may induce compaction of the surface that leads to the production of more runoff. Agricultural practices are probably the single most important factor in affecting erosion rates. For example, Kosmas et al. (1997) demonstrated up to a

twofold difference in erosion under matorral at different rainfall rates, compared to differences of an order of magnitude due to agriculture (although it should be noted that some practices, such as olive cultivation, especially on terraces, can also significantly reduce erosion rates). Concentrated erosion occurs once deeper threads of surface flows are able to overcome the effects of surface cohesion and detach particles rapidly enough to counter refilling of the surface by raindrop-detached

Weathering, Soils, and Slope Processes 5

R

Textural porosity Structural porosity Surface seal R

Interrill sediment yield

4

R

Runoff

3 R

2 R

1 R

0 0

50 Rc (%)

100

Fig. 6.12. Comparison of relative interrill erosion rates as a function of amount and type of rock-fragment cover of the soil surface (modified from Poesen 1992; see also Figure 7 in Poesen et al. 1994).

material (Parsons and Wainwright 2006). This process generally requires flow to enter the turbulent regime. The initiation of rilling will thus be typically during more extreme rainfall events (Figure 6.14), especially if they occur at times of low vegetation cover (Wainwright 1996c; Martínez Casasnovas et al. 2002). Factors that decrease soil cohesion include increased soilmoisture content, coarser textures, low root densities, and high soil dispersivity. Because concentrated flows possess greater energies for detachment, and produce travel distances typically up to tens of metres even for gravel-sized particles, rill-erosion rates are typically much higher than interrill-erosion rates. For example, Bagarello and Ferro (2004) found rill erosion to be responsible for 92.9–99.9 per cent of monitored erosion on plots in southern Italy. If rills become permanent and develop into gullies, then the latter can be responsible for even higher rates of erosion. Poesen et al. (1998) estimated that 80–83 per cent of erosion was due to gully erosion in two areas in south-east Portugal and southeast Spain, although the equivalent value estimated by Valcárcel et al. (2003) for north-western Spain was only

185

27 per cent. However, caution must be exercised when extrapolating from such rates. Although the latter value may be different due to different vegetation conditions in a much more humid area (as found also by Cosandey et al. 1986, in the Cévennes in southern France), there is also an important temporal distinction in rates of gully extension. Kosov et al. (1978 cited in Sidorchuk 1999) suggested that a typical gully will have evolved >90 per cent of its length, 85 per cent of its depth, 60 per cent of its area, and 35 per cent of its volume in the first 5 per cent of its lifetime. The temporal scaling of erosion rates is thus clearly non-linear. This effect can be seen further in the study of Wise et al. (1982), who argued on archaeological grounds that some badland gully systems in southern Spain had evolved little since the Bronze Age, and thus the gullies could be responsible for very little current erosion. Over shorter timescales, there are also important differences—at the same Portuguese and Spanish sites mentioned above, gullies were only responsible for 47–51 per cent of total erosion during a single wet year, compared to the much higher values cited above, which are averages over 3–20 years (Poesen et al. 2002). In extreme cases, especially in highly erodible soils, gullying can expand dramatically, producing entire landscapes that are dominated by these features. These badlands are extensively found in southern Spain, both on marls and unconsolidated bedrocks (Figure 6.15), in south-eastern France in the Black Marls of the Alps, in Italy on marine clays, and in northern Greece (see discussion in Chapter 1). Essentially, badlands relate to the extension of gullies until their contributing areas are reduced to be unable to maintain concentrated overland flows. There are important feedbacks to vegetation growth, with the highly eroded substrates of the badlands being unable to support vegetation due to the low water-holding capacity, high runoff rates, and low nutrient availability. Subsurface erosion can also be an important phenomenon in the Mediterranean, especially in areas of unconsolidated sediments. Soil pipes may occur just beneath the surface, especially if sediments remain disaggregated beneath a crusted surface, allowing them to be detached by vertical and lateral subsurface flows. Piping can also arise due to dissolution, for example in soils or sediments where gypsum or calcium carbonates are present, and may also occur at depth (Harvey 1982). Soil macropores (due to decayed roots or animal burrows, for example) or deep desiccation cracks can allow water to infiltrate rapidly to depth, eroding the macropore or crack as it moves. Such flows may also concentrate along lithological boundaries or structural fissures. Tunnel systems can typically be very

186

John Wainwright

Fig. 6.13. Aerial photograph of Stipa tenacissima vegetated slopes at the Rambla Honda in southern Spain, showing different patterns of connectivity between bare patches (light coloured) in different parts of the catchment, resulting in significant increases in erosion rates in areas where bare patches are interconnected (see Puigdefabregas et al. 1999). Image courtesy of Juan Puigdefabregas.

extensive in susceptible areas (Figure 6.16) and be responsible for rapid transfers of water and sediment into the channel system, often even before main flows across the surface have begun (Farifteh and Soeters 1999). Continued pipe erosion tends to be unstable, leading to subsidence and collapse of the ground surface. Gutierrez-Santolalla et al. (2005) demonstrate the effects of extensive pipes and consequent hazards from gypsum dissolution in the Ebro basin. Pipe collapse is a common cause of rill initiation—Faulkner et al. (2004) discuss the process in dispersive soils in southern Spain—as well as gully initiation (Haigh 1990; Imeson et al. 1982; Martin Penela 1994), not least because collapse can be related to the presence of significant amounts of flowing water under turbulent conditions (Figure 6.17).

While some effort to estimate the relative importance of different water-erosion processes is made above, this chapter does not present a summary table of different measured rates (q.v. Poesen and Hooke 1997 and Table 6.2 in Wainwright and Thornes 2003). The reason for this omission is due to problems that occur when trying to compare erosion rates that are measured at different spatial scales. Parsons et al. (2004) discuss the spatial scaling of erosion rates in detail and show that rates that are measured as specific yields (or areal averages, i.e. in units of M L−2 T−1 ) produce meaningless results. For example, the median rate of the erosion rates in Table 6.2 of Wainwright and Thornes (2003) when expressed in terms of ground lowering is 69.3 mm ka−1 . This rate is in excess of most of the soil-production rates

Weathering, Soils, and Slope Processes

187

Fig. 6.14. Rills formed during the extreme rainfalls of 22 September 1992 in south-eastern France (photo: John Wainwright).

as discussed above, and implies that rather than there being thin soils throughout the Mediterranean, there should in fact be none at all! Although it can be argued that high values in the data relate to human activity,

rates reported under evergreen oak forest in north-east Spain (Sala and Calvo 1990) equate to ground-lowering rates of 156.31 mm ka−1 , while Diamantopoulos (1993) reports rates equivalent to between 19.25 mm ka−1 and

Fig. 6.15. Badlands at Tabernas, Almería, Spain (photo: John Wainwright).

188

John Wainwright

Pipes Drainage network

N 0

500 metres

Agri Valley

Fig. 6.16. Map of observed pipes showing the extent of subsurface erosion in the Agri basin, southern Italy (modified from Farifteh and Soeters 1999).

167.86 mm ka−1 under maquis vegetation in Greece, so this argument does not necessarily always hold. Parsons et al. (2004) demonstrate that the fallacy with the use of specific yields is the assumption that sediment, once detached, remains in transport; discussion of transport distances above suggests that most sediment, especially in unconcentrated flows, travels only a very short distance. Specific yields thus drastically overestimate movement because of redeposition within slope units. The detailed analysis of Martínez Casasnovas

et al. (2002) shows this point very clearly, with extensive areas of redeposition mapped even during an extreme storm event (Figure 6.18).

Processes of Slope Failure and Mass Movement Slope failure occurs when the gravitational force acting in a downslope direction on a mass of soil or rock exceeds the resistance of that mass to movement. The

Weathering, Soils, and Slope Processes

189

Fig. 6.17. Linked pipe and gully erosion, Murcía, Spain (photo: John Wainwright).

corresponding mass movements cover a wide range of styles and size of movement of material. Shallow translational slides or rock topples may involve less than a cubic metre to several cubic metres of material. At the other end of the extreme, large rock avalanches and slides or mudflows may involve even cubic kilometres of material (Figure 6.19). Characterization and definition of different types of movement are described in detail by Dikau et al. (1996). Guzzetti et al. (2002) have investigated the pattern of landslides in relation to the size of movement in areas of central Italy and demonstrated that distributions of landslides triggered by different mechanisms or by single events seem to follow similar patterns of area-frequency relationships (Figure 6.20), which they relate to models of selforganized criticality. Mass movements are a very common process in the Mediterranean and submarine mass movements have also been widely reported (Lasras et al. 2004; Chapter 1). van Asch (1986) suggests that the terrestrial environment is particularly prone to the phenomenon because of the lithology of surface materials, recent and ongoing

seismic activity, climate conditions, and the nature of vegetation cover, all accentuated by the general steepness of slopes. Lithologically, clays commonly occur in soils and weakly consolidated bedrock, as discussed above. Recent sediments are often weakly, if at all, lithified, which further reduces their strength. Elsewhere, tectonic activity has produced significant amounts of fractures and faults, which can act as planes of weakness (Mariolakis 1991). Overfolding in the high mountains often produces the emplacement of rocks above marls and clays, which leads to the possibility of large slides. These areas also contain relict periglacial and glacial deposits above impermeable horizons, further accentuating movement. Uplift also leads in general to oversteepening of slopes (especially when coupled with the removal of sediments at the base by high energy river channels) and thus failure (e.g. Thornes and Alcántara Ayala 1998) (Figure 6.21). Seismic activity also leads to the accentuation of the gravitational force on slopes, and is responsible for a significant number of landslides in the region. Shallow landslides tend to occur as a result of high intensity rainfall events (e.g.

190

John Wainwright

Hillside ditch Vine row

Altitude differences (m)

-0.4 – -0.3 -0.3 – -0.2 –0.2 – –0.1

erosion

–0.1 – 0 0 – 0.1 0.1 – 0.2 0.2 – 0.3

deposition

0.3 – 0.4 0.4 – 0.5

0

N

50 metres

Fig. 6.18. Map of erosion (negative altitude differences) and deposition (positive altitude differences) following an extreme storm event of 215 mm (205 mm falling in 2.25 h, with an average intensity of 91.8 mm h−1 , and a maximum 30-min intensity of 170 mm h−1 ) in Catalonia, Spain in June 2000 (modified from Martínez Casasnovas et al. 2002). The detailed patterns of redeposition within the field demonstrate problems with standard approaches of measuring erosion rates as specific yields (see Parsons et al. 2004).

Wainwright 1996c), but longer periods of rainfall— typically in the winter months—can trigger much deeper-seated failures (e.g. Capecchi and Focardi 1988; Flageollet et al. 1999; Maquaire et al. 2003). Snowmelt is also important at higher altitudes—as seen by the 4,233 landslides mapped by Guzzetti et al. (2002) as a result of a single rapid snowmelt event in Umbria in January 1997 (Figure 6.20). The sparseness of vegetation cover on slopes can also increase the risk of shallow failures, especially on agricultural or sparse matorral areas (van Asch 1980), although the factor is irrelevant for the larger failures, where roots have little or no effect on cohesion of the surface. In some cases, dense vegetation can also lead to failure, by overloading slopes. There is also a possible positive feedback in the occurrence of landslides. Once activated, slope materials can become weakened, and thus slope failure more easily recurs in the same location. Maquaire et al. (2003) investigated several large earthflows in the Barcelonette basin in the French Alps, and suggested that once triggered, surface materials would be likely to remain weakened for up to

150 years (Figure 6.22). This weakening would also produce feedbacks to other processes, such as the lowering of thresholds for the initiation of gullying. Mass movements are responsible for a significant amount of sediment mobilization in the region (e.g. Keefer 1994). Ergenzinger (1992) carried out a detailed analysis of frequency–magnitude relationships for a basin in Calabria and showed that fluvial activity in this basin is controlled by inputs from major, earthquakeinduced landslide events. It can be argued that major landslides are, at least in the longer term, likely to be the most important landscape-modifying process in the Mediterranean landscape.

Slope Hazards Notwithstanding the caveats mentioned above about the extrapolation of soil-erosion rates from small-scale data, it is inescapable that the Mediterranean region has undergone significant periods of soil loss in the past and

Weathering, Soils, and Slope Processes

191

Fig. 6.19. Examples of different types of mass movement: (a) shallow translational slide, French Pre-Alps (photo: John Wainwright); and (b) Super-Sauze earthflow, French Alps (photo: Olivier Maquaire).

192

John Wainwright

Fig. 6.20. Landslide inventories for areas in central Italy: (a) 16,809 landslides in Umbria–Marche (data set A) identified using aerial photographs taken between 1954 and 1956; (b) 4,233 landslides triggered by rapid snow-melt in January 1997 in Umbria (data set B); and (c) non-cumulative frequency–area distributions of these data showing a common slope of –2.5, suggesting characteristics of self-organized criticality (modified from Guzzetti et al. 2002).

under present conditions. Acceleration of soil erosion is often related to human activity, in particular the clearance of vegetation for agriculture. Van Andel and Runnels (1987) discuss a number of events of this nature,

and Wainwright (1994) considers accelerated erosion to be a major factor in population decline in southwest France in the Bronze Age. The development and consequences of this sort of process are considered in

Weathering, Soils, and Slope Processes

193

Fig. 6.21. Large rock slides in the valley of the River Guadalfeo, Granada, Spain, showing the relationship between slope instability and channel undercutting (photo: John Wainwright).

detail by Wainwright and Thornes (2003). As well as on-site impacts in terms of the impoverishment of soil resources and thus long-term land degradation, there are important off-site consequences, such as pollution of watercourses, flood damage, and reservoir siltation (Meddi and Morsli 2001; De Vente and Poesen 2005) (see Chapters 8 and 20).

Guzzetti (2000) has estimated that from 1410 to 1999, at least 840 landslide events were responsible for 10,555 deaths in Italy (Figure 6.23). The sequence is made up of numerous small events punctuated by a number of major events. For example, 400 people died as a result of the failure of a landslide dam and debris flow in the Passer valley in 1419; 1,200 perished as a

194

John Wainwright

(a)

USD LSD

Strength (cohesion, KPa; angle of friction)

First time failure Earthflow reactivation

30

t f p

POC SAU f

VAL

t r

20 t c p 10

USD LSD

t c r

0 0

50

100

150

Time (years) (b) stable Groundwater table fluctuations (due to rainfall and snowmelt)

Factor of safety (Fs)

Long-term increase in Fs due to slope angle reduction and residual strength regain (increase in angle of friction) Decrease in Fs due to strength reduction (peak to residual)

1 unstable S

S

S

S

S

S

S

S

S

S

result of a rockslide at Piuro in 1618; 1,917 were killed by the Vaiont event of 1963 (although some sources say the figure was closer to 2,500). The latter event was the endpoint of a complex series of events following the construction of a dam in the Piave valley, Veneto (Petley 1996). The site is a deep gorge, where detailed mapping before the event revealed a number of older slides, although most of these were not considered to be in danger of reactivation (Semeza and Ghirotti 2000). As the dam construction progressed and reservoir levels started to rise to 590 m a.s.l., there was a minor landslip in March 1960 (Figure 6.24), followed by the onset of creep of slope material in the following months as the reservoir gradually filled. Major fractures occurred within the slope material as a result of loading from the water, leading to a major failure producing 700,000 m3 of sediment in November 1960, with waves of up to 20 m in height being produced against the dam. As a result, the water level in the reservoir was carefully lowered from

Fig. 6.22. Idealized evolution of (a) ground-surface properties and (b) surface instability (factor of safety) based on analysis of earthflows in the Barcelonette basin, France (modified from Maquaire et al. 2003), showing the likelihood of periodic reactivation after failure for a period of up to about 150 years.

the failure height of 650 m to a new level of 600 m a.s.l. in January 1961. It was hoped that creep would produce stabilization of the slope material, and a bypass tunnel was also constructed to help drain the slopes. Following this construction in October 1961, reservoir levels were increased again, reaching 700 m a.s.l. in December 1962. Slope movement increased rapidly to 15 mm day−1 at this stage, so water levels were again reduced to 650 m, which again stabilized movement. It was believed that slope materials had become stable, so water levels were increased once again in April 1963, and slow movements started to occur only at water levels of around 700 m, which appeared to support the idea. Rapid movements started to occur at a height of 710 m a.s.l., but the decision to reduce water levels was delayed, and despite commencing in September 1963, there was a catastrophic slip on 9 October 1963, which displaced 250 M m3 of sediment at velocities up to 30 m s−1 . About 30 M m3 of water was displaced by this sediment,

Weathering, Soils, and Slope Processes

195

(a) Vajont Oct.9 1963

2

10

Number of dead and missing people per year (thousands)

Passer Valley Sept. 22 1419

0 1400

1500

5

Montepiano June 24 1764

1

1600

1700

1800

1900

0 2000

(b) Vajont Oct. 9 1963

2

1 Cetera Oct. 24 1910

0 1900

Sarno May 5 1998

Salerno Oct. 25 1924 M. Antelao Feb. 17 1925

1920

10

1940

Salerno Oct. 25 1954

1960

5

Cumulative number of dead and missing people per year (thousands)

Piuro Sept. 4 1618

Stava July 19 1985

1980

overtopping the reservoir with waves reaching up to 950 m a.s.l., with consequent flooding of downstream settlements leading to the fatalities noted above. Semeza and Ghirotti (2000) emphasize the need for detailed mapping of local conditions, including the presence of older failures, and indeed landslide-susceptibility mapping is an increasingly common planning tool in the minimization of landslide hazards (e.g. Carrara et al. 1991; Gokceoglu et al. 2005). Detailed monitoring using GPS and ground-based techniques is also becoming important, as shown in the assessment of hazards in the Super Sauze earthflows in the French Alps (Maquaire et al. 2003).

The Wider Context In Chapter 1, the link between uplift and erosion has already been touched upon. It is important to recognize

0 2000

Fig. 6.23. Analysis of landslide events that resulted in fatalities in Italy comparing (a) the period 1410–1999 and (b) 1900–99 (modified from Guzzetti 2000).

that there is a complex interplay between uplift, topography, and erosion rates (Burbank and Anderson 2001). Lewis et al. (2000) have suggested that the continued uplift (despite crustal extension) of the Catalan coastal ranges is as a result of flexural isostasy caused by erosion of the rifting margin. Increased topography is likely to accelerate erosion rates in this case, producing a positive feedback. Gargani (2004) has argued that the topography of the Rhône valley in southern France is a result of complex isostatic compensation during the Messinian Salinity Crisis due to unloading of the seabed as sea levels dropped, and due to the consequent incision of the deep river gorge. At a larger scale, Kuhlemann et al. (2002) estimate that erosion of the Alps following uplift has produced 925,800 km3 of sediment over the last 35 Ma. Isostatic compensation for removal of material causes continued uplift, and these figures suggest that almost 7 km of uplift may be the result of the isostatic compensation from this erosion. Similarly, the Rhône valley

(a)

(b ) Before 9 October 1963

Pontesei Reservoir

m 2000 M ae

Mt. Toc

1500 Mt. Salta Longaron Longaronee

Vajont Ex Vajont Reservoir

2 Km

Vajont Valley

1000

ITALY

0

Mt. Salta

1963 Vajont lake

500

Mt. Toc

After 9 October 1963 Quaternary/Stratified alluvial gravels Upper Cretaceous–Lower Palaeocene (Scaglia Rossa) Cretaceous–Jurrasic formations (Socchér formation sensulato and coeval. Socchér formation sensu stricto / Ammonitico rosso and Fonzaso formation) Dogger (Calcare del Vajont)

m 2000 1500 1000 500

Upper Liassic (lgne formation) Lower and Middle Liassic (Soverzene formation)

(c ) Precipitation (10 day increments)

Upper Triassic (Dolomia principale) Faults and overthrusts

mm

300

Landslide failure surfaces Direction of waterflow into aquifers

0 Reservoir level m

720

710 m Perimeter crack March 1960 slide

November 1960 slide

October 1963 slide

Construction of by-pass tunnel

580 Rate of movement cm/day

4

0 Water level in piezometers m

720

710 m P2

P1

P3

620 1960

1961

1962

1963

Fig. 6.24. Details of the Vaiont landslide disaster: (a) location map; (b) geological cross-sections before and after the major event of 9 October 1963; and (c) comparisons of precipitation, reservoir level, rate of movement, and subsurface water levels during the construction, leading up to the failure event (modified from Semenza and Ghirotti 2000).

Weathering, Soils, and Slope Processes

will have subsided in compensation for the 30,000 km3 deposited over this time period, while deposition in the Gulf of Lions is of the order of 267,000 km3 and in the Adriatic, of the order of 104,700 km3 , with subsidence in both cases having important feedbacks to coastal and fluvial processes. It has also been argued that the relative rates of plate convergence compared to erosion is an important factor in controlling the shape of orogens (Willett et al. 2001), while Whipple and Meade (2006) have argued that erosion rates can even control plate movements by means of these feedbacks. Weathering and soil erosion can also have major impacts on the carbon cycle. Geologically, these processes can feed back to large-scale climate changes (as discussed for the Himalaya, Tibet, and Andes by Raymo et al. 1988, who argued that uplift would increase weathering rates, leading to a drawdown of atmospheric CO2 and thus to global-scale cooling). Lal (2005) has argued that the state of knowledge is presently incomplete as to whether soil erosion acts as a sink or a source for carbon, with important consequences for ongoing climate change. However, estimates that burial of eroded sediments on land can explain the missing carbon from the global cycle (Stallard 1998) is based on an erroneous assumption about the extent of such burial, related to the issues of scaling erosion rates already mentioned above (Parsons et al. 2004). Alternative hypotheses suggest that erosion ultimately provides a net source of carbon to the atmosphere, although Lal (2003) makes the same erroneous assumption as Stallard in the amount of total erosion that can take place. Whichever of these theories proves to be correct, the Mediterranean as a dynamically eroding area is likely to be a major area of concern for global carbon cycles—for example, Roose (2004) summarized data for carbon erosion from sites in Algeria and other semi-arid areas, and suggested that yields of carbon could be significant, especially on agricultural soils and on sparsely vegetated, steep slopes–and is potentially a key area for attempting to mitigate climate-change impacts in a holistic way.

Conclusion Several decades of research employing a processbased perspective have produced important advances in the understanding of the operation of hillslopes in the Mediterranean region. It is important to recognize the variability inherent within these systems and that more detailed understanding allows the potential for appropriate management of these environments. This vari-

197

ability implies that perspectives that are characterized by ‘typical Mediterranean’ conditions are likely to lead to problems, and that a bottom-up (local) approach is required to understanding patterns of slope processes, be they weathering, water erosion, or mass movements. The reflection of different timescales in processes highlights the vulnerability of these systems, especially the relatively slow rate of soil production compared to erosion rates, notwithstanding the problems highlighted with the measurement of the latter. What is now required is a more holistic perspective, in which information about these different processes is integrated—including over longer timescales—and issues of compatibility addressed. This chapter has highlighted a number of (but by no means all) important non-linearities and scaling issues within and between different processes. The evaluation of these non-linearities and their impacts must remain a priority in the investigation of Mediterranean landforms. They are significant in understanding the evolution of the Mediterranean, and will become increasingly important in the sustainable management of these sensitive environments, with regional and global implications.

References Aberg, G., Charalampides, G., Fosse, G., and Hjelmseth, H. (2001), The use of Pb isotopes to differentiate between contemporary and ancient sources of pollution in Greece, Atmospheric Environment 35: 4609–15. Ahnert, F. (1976), Brief description of a comprehensive threedimensional model of landform development, Zeitschrift für Geomorphologie NS Suppl. 25: 29–49. Arianoutsou, M. (1993), Leaf litter decomposition and nutrient release in a maquis (evergreen sclerophyllous) ecosystem of north-eastern Greece, Pedobiologia 37: 65–71. Bagarello, V. and Ferro, V. (2004), Plot-scale measurement of soil erosion at the experimental area of Sparacia (southern Italy), Hydrological Processes 18: 141–57. Ballais, J.-L. and Bosc, M. C. (1994), The ignifracts of the SainteVictoire Mountain (Lower Provence, France), in M. Sala and J. L. Rubio (eds.), Soil Erosion and Degradation as a Consequence of Forest Fires. Geoforma Ediciones, Logroño, 217–27. Birkeland, P. W. (1999), Soils and Geomorphology. 3rd edn. Oxford University Press, Oxford. Bochet, E., Poesen, J., and Rubio, J. L. (2000), Mound development as an interaction of individual plants with soil, water erosion and sedimentation processes on slopes, Earth Surface Processes and Landforms 25: 847–67. (2002), Influence of plant morphology on splash erosion in a Mediterranean matorral, Zeitschrift für Geomorphologie 46: 223–43. Boero, V. and Schwertmann, U. (1989), Occurrence and transformations of iron and manganese in a colluvial terra rossa toposequence of northern Italy, Catena, 14: 519–31. Bolle, H. J. (1995), Climate and desertification, in R. Fantechi, D. Peter, P. Balabanis, and J. L. Rubio (eds.), Desertification in a

198

John Wainwright

European Context: Physical and Socio-Economic Aspects, EC, Luxembourg, 15–34. Bücher, A. (1989), Fallout of Saharan dust in the Northwestern Mediterranean region, in M. Leinen and M. Sarnthein (eds.), Palaeoclimatology and Palaeometeorology: Modern and Past Patterns of Global Atmospheric Transport. Kluwer Academic, Dordrecht, 565–84. Bufalo, M., Oliveros, C., and Quélennec, R. E. (1989), L’érosion des Terres Noires dans la région du Buëch (Hautes-Alpes). Contribution à l’étude des processus érosifs sur le bassin versant représentatif (BVRE) de Saint-Genis, La Houille Blanche 3/4: 193–5. Burbank, D. W. and Anderson, R. S. (2001), Tectonic Geomorphology. Blackwell Science, Oxford. Calaforra, J. M. and Pulido-Bosch, A. (2003), Evolution of the gypsum karst of Sorbas (SE Spain), Geomorphology 50: 173–80. Cammeraat, E. L. H. (2002), A review of two strongly contrasting geomorphological systems within the context of scale, Earth Surface Processes and Landforms 27: 1201–22. (2004), Scale dependent thresholds in hydrological and erosion response of a semi-arid catchment in southeast Spain, Agriculture Ecosystems and Environment 104: 317–32. Capecchi, F. and Focardi, P. (1988), Rainfall and landslides: research into a critical precipitation coefficient in an area of Italy, in C. Bonnard (ed.), Landslides, A. A. Balkema, Rotterdam, 1131–6. Carrara, A., Cardinali, M., Detti, R., Guzzetti, F., Pasqui, V., and Reichenbach, P. (1991), GIS techniques and statisticalmodels in evaluating landslide hazard, Earth Surface Processes and Landforms 16: 427–45. Casals, P., Romanyà, J., Cortina, J., Bottner, P., Coûteaux, M.-M., and Vallejo, V. R. (2000), CO2 efflux from a Mediterranean semi-arid forest soil, I. Seasonality and effects of stoniness, Biogeochemistry 48: 261–81. Charley, J. L. and West, N. E. (1975), Plant-induced soil-chemical patterns in some shrub-dominated semi-desert ecosystems of Utah, Journal of Ecology 63: 945–64. Correggiari, A., Guerzoni, S., Lenaz, R., Quarantotto, G., and Rampazzo, G. (1989), Dust deposition in the central Mediterranean (Tyrrhenian and Adriatic Seas): relationships with marine sediments and riverine input, Terra Nova 1: 549–58. Cosandey, C., Billard, A., and Muxart, T. (1986), Present day evolution of gullies formed in historical times in the Montagne du Lingas, Southern Cévennes, France, in V. Gardiner (ed.), International Geomorphology. John Wiley & Sons, Chichester, ii. 523–31. Dana, E. D. and Mota, J. F. (2006), Vegetation and soil recovery on gypsum outcrops in semi-arid Spain, Journal of Arid Environments 65: 444–59. Davies, F. B. M. and Notcutt, G. (1988), Accumulation of fluoride by lichens in the vicinity of Etna volcano, Water, Air and Soil Pollution 42: 365–71. De Alba, S., Borselli, L., Torri, D., Pellegrini, S., and Bazzoffi, P. (2006), Assessment of tillage erosion by mouldboard plough in Tuscany (Italy), Soil and Tillage Research 85: 123–42. Vente, J. and Poesen, J. (2005), Predicting soil erosion and sediment yield at the basin scale: Scale issues and semi-quantitative models, Earth-Science Reviews 71: 95–125. Bazzoffi, P., Van Rompaey, A., and Verstraeten, G. (2006), Predicting catchment sediment yield in Mediterranean environments: the importance of sediment sources and connectivity in Italian drainage basins, Earth Surface Processes and Landforms 31: 1017–34.

Dercourt, J., Zonenshain, L. P, Ricou, L.-E., Kazmin, V. G., Le Pichon, X., Knipper, A. L., Grandjacquet, C., Sbortshikov, I. M., Geyssant, J., Lepvrier, C., Pechersky, D. H., Boulin, J., Sibuet, J.-C., Savostin, L. A., Sorokhtin, O., Westphal, M., Bazhenov, M. L., Lauer, J. P., and Biju-Duval, B. (1986), Geological evolution of the Tethys belt from the Atlantic to the Pamirs since the Lias, Tectonophysics 123: 241–315. Dewey, J. F., Helman, M. L., Turco, E., Hutton, D. H. W., and Knott, S. D. (1989), Kinematics of the western Mediterranean in M. P. Coward, D. Dietrich, and R.G. Park (eds.), Alpine Tectonics. Geological Society Special Publication 45, London, 265–83. Diamantopoulos, J. (1993), Field site investigations: Petralona, Thessaloniki, Greece, in J. B. Thornes (ed.), MEDALUS I Final Report, Final Report on Contract EPOC-CT90-0014-(SMA), EU, Brussels, 560–80. Dikau, R., Brunsden, D., Schrott, L., and Ibsen, M. L. (eds.) (1996), Landslide Recognition. Identification, Movement and Causes. John Wiley & Sons, Chichester. Do˘gan, U. and Özel, S. (2005), Gypsum karst and its evolution east of Hafik (Sivas, Turkey), Geomorphology 71: 373–88. Dreybrodt, W. (1988), Processes in Karst Systems. Springer, Berlin. Durand, P., Neal, C., Lelong, F., and Didonlescot, J. F. (1992), Effects of land use and atmospheric input on stream and soil chemistry—field results and long-term simulation at MontLozère (Cévennes National Park, Southern France, Science of the Total Environment 119: 191–209. Ergenzinger, P. (1992), A conceptual geomorphological model for the development of a Mediterranean river basin under neotectonic stress (Buonamico basin, Calabria, Italy), in D. E. Walling, T. R. Davies, and B. Hasholt (eds.), Erosion, Debris Flows and Environment in Mountain Regions. IAHS Publication 209, IAHS, Wallingford, 51–60. Farifteh, J. and Soeters, R. (1999), Factors underlying piping in the Basilicata region, southern Italy, Geomorphology 26: 239–51. Faulkner, H., Alexander, R., Teeuw, R., and Zukowsky, P. (2004), Variations in soil dispersivity across a gully head displaying shallow sub-surface pipes, and the role of shallow pipes in rill initiation, Earth Surface Processes and Landforms 29: 1143–60. Fioretto, A., Papa, S., and Fuggi, A. (2003), Litter-fall and litter decomposition in a low Mediterranean shrubland, Biology and Fertility of Soils 39: 37–44. Flageollet, J. C., Maquaire, O., Martin, B., and Weber, D. (1999) Landslides and climatic conditions in the Barcelonnette and Vars basins (Southern French Alps, France), Geomorphology 30: 65–78. Frumkin, A. and Fischhendler, I. (2005), Morphometry and distribution of isolated caves as a guide for phreatic and confined paleohydrological conditions, Geomorphology 67: 457–71. and Stein, M. (2004), The Sahara-East Mediterranean dust and climate connection revealed by strontium and uranium isotopes in a Jerusalem speleothem, Earth and Planetary Science Letters 217: 451–64. Gallart, F., Llorens, P., Latron, J., and Regues, D. (2002), Hydrological processes and their seasonal controls in a small Mediterranean mountain catchment in the Pyrenees, Hydrology and Earth System Sciences 6: 527–37. Gams, I., Nicod, J., Julian, M., Anthony, E., and Sauro, U. (1993), Environmental change and human impacts on the Mediterranean karsts of France, Italy and the Dinaric region, in

Weathering, Soils, and Slope Processes P. W. Williams (ed.), Karst Terrains, Environmental Changes, Human Impact. Catena Suppl. 25, Cremlingen. 59–98. García-Ruiz, J. M., Valero-Garcés, B., González-Sampériz, P., Lorente, A., Martí-Bono, C., Beguería, S., and Edwards, L. (2001), Stratified scree in the Central Spanish Pyrenees: palaeoenvironmental implications’ Permafrost and Periglacial Processes 12: 233–42. Gargani, J. (2004), Modelling of the erosion in the Rhone valley during the Messinian crisis (France), Quaternary International 121: 13–22. Garty, J. (1992), The post-fire recovery of rock-inhabiting algae, microfungi and lichens, Canadian Journal of Botany 70: 301–12. Giovannini, G., Lucchesi, S., and Giachetti, M. (1987), The natural evolution of a burned soil: a three-year investigation, Soil Science 143: 220–6. (1988), Effect of heating on some physical and chemical parameters related to soil aggregation and erodibility, Soil Science 146: 255–61. (1990), Beneficial and detrimental effects of heating on soil quality, in J. G. Goldammer and M. J. Jenkins (eds.), Fires in Ecosystem Dynamics. Proceedings of the Third International Symposium on Fire Ecology, Freiburg, FRG, May 1989, SPB Academic, The Hague, 95–102. Gokceoglu, C., Sonmez, H., Nefeslioglu, H. A., Duman, T. Y., and Can, T. (2005), The 17 March 2005 Kuzulu landslide (Sivas, Turkey) and landslide-susceptibility map of its near vicinity, Engineering Geology 81: 65–83. Gomez, J. A., Romero, P., Giraldez, J. V., and Fereres, E. (2004), Experimental assessment of runoff and soil erosion in an olive grove on a Vertic soil in southern Spain as affected by soil management, Soil Use and Management 20: 426–31. Govers, G., Vandaele, K., Desmet, P. J. J., Poesen, J., and Bunte K. (1994), The role of tillage in soil redistribution on hillslopes, European Journal of Soil Science 45: 469–78. Günster, N. and Skowronek, A. (2001), Sediment-soil sequences in the Granada Basin as evidence for long- and short-term climatic changes during the Pliocene and Quaternary in the Western Mediterranean, Quaternary International 78: 17–32. Gutierrez-Santolalla, F., Gutierrez-Elorza, M., Marin, C., Maldonado, C., and Younger, P. L. (2005), Subsidence hazard avoidance based on geomorphological mapping in the Ebro River valley mantled evaporite karst terrain (NE Spain), Environmental Geology 48: 370–83. Guzzetti, F. (2000), Landslide fatalities and the evaluation of landslide risk in Italy, Engineering Geology 58: 89–107. Malamud, B. D., Turcotte, D. L., and Reichenbach, P. (2002), Power-law correlations of landslide areas in central Italy, Earth and Planetary Science Letters 195: 169–83. Gvirtzman, G. and Wieder, M. (2001), Climate of the last 53,000 years in the eastern Mediterranean, based on soil-sequence stratigraphy in the coastal plain of Israel, Quaternary Science Reviews 20: 1827–49. Haigh, M. J. (1990), Evolution of an anthropogenic desert gully system, in D. E. Walling, A. Yair, and S. Berkowicz (eds.), Erosion, Transport and Deposition Processes (Proceedings of the Jerusalem Workshop, March-April 1987). IAHS Publication 189, IAHS Press, Wallingford, 65–77. Harvey, A. M. (1982), The role of piping in the development of badlands and gully systems in south-east Spain, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. Geobooks, Norwich, 317–35. Hassouni, K., and Bouhlassa, S. (2006), Estimate of soil erosion on cultivated soils using Cs-137 measurements and calibration

199

models: a case study from Nakhla watershed, Morocco, Canadian Journal of Soil Science 86: 77–87. Hughes, P. D., Woodward, J. C., and Gibbard, P. L. (2006), Quaternary glacial history of the Mediterranean mountains, Progress in Physical Geography 30: 334–64. Imeson, A. C., Kwaad, F. J. P. M., and Verstraten, J. M. (1982), The relationship of soil physical and chemical properties to the development of badlands in Morocco, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. GeoBooks, Norwich, 47–70. Verstraten, J. M., van Mulligen, E. J., and Sevink, J. (1992), The effects of fire and water repellency on infiltration and runoff under Mediterranean type forest, Catena 19: 345–61. Issar, A., Tsoar, H., and Levin, D. (1989), Climatic changes in Israel during historical times and their impact on hydrological, pedological and socio-economic systems, in M. Leinen and M. Sarnthein (eds.), Palæoclimatology and Palæometeorology: Modern and Past Patterns of Global Atmospheric Transport. Kluwer Academic, Dordrecht, 359–83. Jackson, M. L., Clayton, R. N., Violante, A., and Violante, P. (1982), Eolian influence on terra rossa soils of Italy traced by quartz oxygen isotope ratios, in H. van Olpen and F. Veniale (eds.), International Clay Conference. Elsevier, Amsterdam, 293–301. Jennings, J. N. (1985), Karst Geomorphology. Blackwell, Oxford. Kaufmann, G. and Braun, J. (2001), Modelling karst denudation on a synthetic landscape, Terra Nova 13: 313–20. Keefer, D. K. (1994), The importance of earthquake-induced landslides to long-term slope erosion and slope-failure hazards in seismically active regions, Geomorphology 10: 265–84. Kirkby M. J., Bracken, L. J., and Reaney, S. (2002), The influence of land use, soils and topography on the delivery of hillslope runoff to channels in SE Spai, Earth Surface Processes and Landforms 27: 1459–73. Kosmas, C., Danalatos, N. G., and Gerontidis, S. (2000), The effect of land parameters on vegetation performance and degree of erosion under Mediterranean conditions, Catena 40: 3–17. Cammeraat, L. H., Chabart, M., Diamantopoulis, J., Farand, R., Gutierrez, L., Jacob, A., Marques, H., MartinezFernandez, J., Mizara, A., Moustakas, N., Nicolau, J. M., Oliveros, C., Pinna, G., Puddu, R., Puigdefábregas, J., Roxo, M., Simao, A., Stamou, G., Tomasi, N., Usai D., and Vacca, A. (1997), The effect of land use on runoff and soil erosion rates under Mediterranean conditions, Catena 29: 45–59. Kuhlemann, J., Frisch, W., Székely, B., Dunkl, I., and Kázmér, M. (2002), Post-collisional sediment budget history of the Alps: tectonic versus climatic control. International Journal of Earth Sciences 91: 818–37. Lal, R. (2003), Soil erosion and the global carbon budget, Environment International 29: 437–50. (2005), Soil erosion and carbon dynamics, Soil and Tillage Research 81: 137–42. Larcheveque, M., Montes, N., Baldy, V., and Dupouyet, S. (2005), ‘Vegetation dynamics after compost amendment in a Mediterranean post-fire ecosystem’, Agriculture Ecosystems and Environment 110: 241–8. Larrasoana, J. C., Roberts, A. P., Rohling, E. J., Winklhofer, M., and Wehausen, R. (2003), Three million years of monsoon variability over the northern Sahara, Climate Dynamics 21: 689–98. Lastras, G., Canals, M., Urgeles, R., De Bastist, M., Calafat, A. M., and Casamor, J. L. (2004), Characterisation of the

200

John Wainwright

recent BIG’95 debris flow deposit on the Ebro margin, western Mediterranean Sea, after a variety of seismic reflection data. Marine Geology 213: 235–55. Lavee, H., Kutiel, P., Segev, M., and Benyamini, Y. (1995), Effect of surface roughness on runoff and erosion in a Mediterranean ecosystem: the role of fire, Geomorphology 11: 227–34. Leone, G., Bonadonna, F., and Zanchetta, G. (2000), Stable isotope record in mollusca and pedogenic carbonate from Late Pliocene soils of Central Italy, Palaeogeography, Palaeoclimatology, Palaeoecology 163: 115–31. Lewis, C. J., Verges, J., and Marzo, M. (2000), High mountains in a zone of extended crust: Insights into the Neogene-Quaternary topographic development of northeastern Iberia, Tectonics 19: 86–102. MacLeod, D. A. (1980), The origin of the red Mediterranean soils in Epirus, Greece, Journal of Soil Science 31: 125–36. Malet J.-P., Auzet, A. V., Maquaire, O., Ambroise, B., Descroix, L., Esteves, M., Vandervaere, J. P., and Truchet, E. (2003), Soil surface characteristics influence on infiltration in black marls: Application to the Super-Sauze earthflow (southern Alps, France), Earth Surface Processes And Landforms 28: 547–64. Maquaire, O., Malet, J.-P., Remaître, A., Locat, J., Klotz, S., and Guillon, J. (2003), Instability conditions of marly hillslopes: towards landsliding or gullying? The case of the Barcelonnette Basin, South East France’ Engineering Geology 70: 109–30. Mariolakos, I. (1991), Prediction of natural mass movement in tectonically active areas, in M. E. Almeida-Teixeira, R. Fantechi, R. Oliveira, and A. Gomes Coelho (eds.), Natural Hazards and Engineering Geology. Prevention and Control of Landslides and Other Mass Movements. Commission of the European Communities, EUR 12918, Brussels, 69–81. Martin Penela, A. J. (1994), Pipe and gully systems— development in the Almanzora basin (Southeast Spain), Zeitschrift für Geomorphologie 38: 207–22. Martínez Casasnovas, J. A., Ramos, M. C., and Ribes Dasi, M. (2002), Soil erosion caused by extreme rainfall events: mapping and quantification in agricultural plots from very detailed digital elevation models, Geoderma 105: 125–40. Martínez Mena, M., Castillo, V., and Albaladejo, J. (2002), Relations between interrill erosion processes and sediment particle size distribution in a semiarid Mediterranean area of SE of Spain, Geomorphology 45: 261–75. Meddi, M. and Morsli, B. (2001), Erosion and runoff in BeniChougrane’s mountains (Western Algeria), Zeitschrift für Geomorphologie 45: 443–52. Mommersteeg, H. J. P. M., Loutre, M. F., Young, R., Wijmstra, T. A., and Hooghiemstra, H. (1995), Orbital forced frequencies in the 975 000 year pollen record from Tenaghi Philippon (Greece), Climate Dynamics 11: 4–24. Moresi, M. and Mongelli, G. (1988), The relation between the terra rossa and the carbonate-free residue of the underlying limestones and dolostones in Apulia, Italy, Clay Minerals 23: 439–46. Naveh, Z. (1975), The evolutionary significance of fire in the Mediterranean region, Vegetatio 29: 199–209. New, M., Lister, D., Hulme, M., and Makin, I. (2000), A highresolution data set of surface climate over global land areas, Climate Research 21: 1–25. Nihlen, T., Mattsson, J. O., Rapp, A., Gagaoudaki, C., Kornaros, G., and Papageorgiou, J. (1995), Monitoring of Saharan dust fallout on Crete and its contribution to soil formation, Tellus Series B—Chemical and Physical Meteorology 47: 365–34.

NOAA (1984) FAO/UNESCO UNEP gridded FAO/UNESCO soil units, , accessed August 2001. Oehm, B. and Hallet, B. (2005), Rates of soil creep, worldwide: weak climatic controls and potential feedback, Zeitschrift für Geomorphologie 49: 353–72. Oliva, P., Dupre, B., Martin, F., and Viers, J. (2004), The role of trace minerals in chemical weathering in a highelevation granitic watershed (Estibere, France): Chemical and mineralogical evidence, Geochimica et Cosmochimica Acta 68: 2223–43. Oostwoud Wijdenes, D. J., Poesen, J., Vandekerckhove, L., and Kosmas, C. (2001), Measurements at one-year interval of rock-fragment fluxes by sheep trampling on degraded rocky slopes in the Mediterranean, Zeitschrift für Geomorphologie 45: 477–500. Ortiz, I., Simón, M., Dorronsoro, C., Martín, F., and García, I. (2002), Soil evolution over the Quaternary period in a Mediterranean climate (SE Spain), Catena 48: 131–48. Palacios, D., de Andrés, N., and Luengo, E. (2003), Distribution and effectiveness of nivation in Mediterranean mountains: Peñalara (Spain), Geomorphology 54: 157–78. Parsons, A J. and Wainwright, J. (2006), Depth distribution of interill overland flow and the formation of rills, Hydrological Processes 20: 1511–23. Abrahams, A. D., and Simanton, J. D. (1992), Microtopography and soil-surface materials on semiarid piedmont hillslopes, southern Arizona, Journal of Arid Environments 22: 107–15. Wainwright, J., Powell, D. M., Kaduk, J., and Brazier, J. E. (2004), A conceptual model for understanding and predicting erosion by water, Earth Surface Processes and Landforms 29: 1293–302. Peltier, L. C. (1950), The geographic cycle in periglacial regions as it is related to climatic geomorphology, Annals of the Association of American Geographers 40: 214–36. Petley, D. (1996), The mechanics and landforms of deep-seated landslides, in M. G. Anderson and S. M. Brooks (eds.), Advances in Hillslope Processes. John Wiley & Sons, Chichester, ii. 826–36. Poesen, J. W. A. (1992), Mechanisms of overland flow generation and sediment production on loamy and sandy soils with and without rock fragments, in A. J. Parsons and A. D. Abrahams (eds.), Overland Flow Hydraulics and Erosion Mechanics. UCL, London, 275–305. Torri, D., and Bunte, K. (1994), Effects of rock fragments on soil erosion by water at different spatial scales—a review. Catena 23: 141–66. and Hooke, J. M. (1997), Erosion, flooding and channel management in Mediterranean environments of southern Europe, Progress in Physical Geography 21: 157–99. Vandaele, K., and van Wesemael, B. (1998), Gully erosion: importance and model implication, in J. Boardman and D. Favis-Mortlock (eds.), Modelling Soil Erosion by Water. Springer, Berlin, 285–311. Bunte, K., and Solé Benet, A. (1998), Variation of rock fragment cover and size along semiarid hillslopes: a case-study from southeast Spain, Geomorphology 23: 323–35. Vandekerckhove, L., Nachtergaele, J., Oostwoud Wijdenes, D., Verstraeten, G., and van Wesemael, B. (2002), Gully erosion in dryland environments, in L. J. Bull and M. J. Kirkby (eds.), Dryland Rivers: Hydrology and Geomorphology of SemiArid Channels. John Wiley & Sons, Chichester, 229–62.

Weathering, Soils, and Slope Processes Pope, G. A., Dorn, R. I., and Dixon, J. C. (1995), A new conceptual model for understanding geographical variations in weathering, Annals of the Association of American Geographers 85: 38– 64. Puigdefabregas, J., Sole, A., Gutierrez, L., del Barrio, G., and Boer, M. (1999), Scales and processes of water and sediment redistribution in drylands: results from the Rambla Honda field site in Southeast Spain, Earth-Science Reviews 48: 39–70. Pulido-Villena, E., Reche, I., and Morales-Baquero, R. (2006), Significance of atmospheric inputs of calcium over the southwestern Mediterranean region: high mountain lakes as tools for detection, Global Biogeochemical Cycles 20/2: GB2012. Pye, K. (1992), ‘Aeolian dust transport and deposition over Crete and adjacent parts of the Mediterranean Sea’, Earth Surface Processes and Landforms 17: 271–88. Raymo, M. E., Ruddiman, W. F., and Froelich, P. N. (1988), Influence of late Cenozoic mountain building on ocean geochemical cycles, Geology 16: 649–53. Roose, E. (2004), Erosion du carbone et indice de sélectivité à l’échelle de la parcelle dans les régions tropicales et méditerranéennes, in E. Roose, G. de Noni, C. Prat, F. Ganry, and G. Bourgeon (eds.), Gestion de la Biomasse, Erosion et Séquestration du Carbone. Actes—Erosion du Carbone (Colloque International au Centre Agropolis, Montpellier, 23–28/9/2002). Bulletin du Reseau Erosion 22, Montpellier, 74–94. Sáez, A., Inglès, M., Cabrera, L., and de las Heras, A. (2003), Tectonic–palaeoenvironmental forcing of clay-mineral assemblages in nonmarine settings: the Oligocene–Miocene As Pontes Basin (Spain), Sedimentary Geology 159: 305–24. Sala, M. and Calvo, A. (1990), Response of four different Mediterranean vegetation types to runoff and erosion, in J. B. Thornes (ed.), Vegetation and Erosion. John Wiley & Sons, Chichester, 347–62. Sanjurjo, M. J. F., Corti, G., Certini, G., and Ugolini, F. C. (2003), Pedogenesis induced by Genista aetnensis (Biv.) DC. on basaltic pyroclastic deposits at different altitudes, Mt. Etna, Italy, Geoderma 115: 223–43. Savat, J. (1981), Work done by splash—laboratory experiments, Earth Surface Processes and Landforms 6: 275–83. Scarciglia, F., Le Pera, E., and Critelli, S. (2005), Weathering and pedogenesis in the Sila Grande Massif (Calabria, South Italy): from field scale to micromorphology, Catena 61: 1–29. Schulte, L. and Julia, R. (2001), ‘Quaternary soil chronosequences of southeastern Spain, Zeitschrift für Geomorphologie 45: 145–58. Semenza, E. and Ghirotti, M. (2000), History of the 1963 Vaiont slide: the importance of geological factors, Bulletin of Engineering Geology and the Environment 5: 87–97. Serrat, P. (1999), Dynamique sédimentaire actuelle d’un système fluvial méditerranéen: l’Agly (France), Comptes Rendus de l’Académie des Sciences—Series IIA—Earth and Planetary Science 329: 189–96. Shakesby, R. A. and Doerr, S. H. (2006), Wildfire as a hydrological and geomorphological agent, Earth-Science Reviews 74,: 269–307. Coelho, C. D. A., Ferreira, A. D., Terry, J. P., and Walsh, R. P. D. (1993), Wildfire impacts on soil-erosion and hydrology in wet Mediterranean forest, Portugal, International Journal of Wildland Fire 3: 95–110. Sidorchuk, A. (1999), Dynamic and static models of gully erosion, Catena 37: 401–14. Sorriso Valvo, M. and Sylvester, A. G. (1993), The relationship between geology and landforms along a coastal mountain

201

front, northern Calabria, Italy, Earth Surface Processes and Landforms 18: 257–73. Stallard, R. F. (1998), Terrestrial sedimentation and the carbon cycle: Coupling weathering and erosion to carbon burial, Global Biogeochemical Cycles 12: 231–57. Sumner, M. E. and Wilding, L. P. (2000), Introduction, in M. E. Sumner (ed.), Handbook of Soil Science. CRC, Boca Raton, pp. xvii–xx. Thomas, D. S. G. (1997), Arid environments: their nature and extent, in D. S. G. Thomas (ed.), Arid Zone Geomorphology. Process, Form and Change in Drylands. John Wiley & Sons, Chichester, 1–12. Thornes, J. B. (1985), The ecology of erosion, Geography 70: 222–36. (1990), The interaction of erosional and vegetational dynamics in land degradation: spatial outcomes, in J. B. Thornes (ed.), Vegetation and Erosion. John Wiley & Sons, Chichester, 41–53. and Alcántara Ayala, I. (1998), Modelling mass failure in a Mediterranean mountain environment: climatic, geological, topographical and erosional controls, Geomorphology 24: 87–100. Tomadin, L., Lenaz, R., Landuzzi, V., Mazzucotelli, A., and Vannucci, R. (1984), Wind-blown dusts over the central Mediterranean, Oceanologica Acta 7: 13–23. and Sartori, R. (1990), Clay minerals as natural tracers in sediments, water column, and lower atmosphere of the Tyrrhenian basin (Mediterranean Sea), Mineralogica Petrologica Acta 33: 81–91. Torri, D. and Poesen, J. W. A. (1992), The effect of soil surface slope on raindrop detachment, Catena 19: 561–78. Sfalanga, M., and del Sette, M. (1981), Splash detachment: runoff depth and soil cohesion, Catena 14: 149–55. Valcárcel, M., Taboada, M. T., Paz, A., and Dafonte, J. (2003), Ephemeral gully erosion in Northwestern Spain, Catena 50: 199–216. van Andel, T. H. and Runnels, C. N. (1987), Beyond the Acropolis. A Rural Greek Past. Stanford University Press, Stanford. Zangger, E., and Demitrack, A. (1990), Land use and soil erosion in Prehistoric and Historical Greece. Journal of Field Archaeology 17: 379–96. van Asch, T. W. J. (1980), Water erosion on slopes and landsliding in a Mediterranean landscape, Utrechtse Geografische Studies 20. (1986), Hazard mapping as a tool for landslide prevention in Mediterranean areas, in R. Fantechi and N. S. Margaris (eds.), Desertification in Europe. Reidel, Dordrecht, 126–35. van Beek, L. P. H., Wint, J., Cammeraat, L. H., and Edwards, J. P. (2005), Observation and simulation of root reinforcement on abandoned Mediterranean slopes, Plant and Soil 278: 55–74. van Wesemael, B. (1993), Litter decomposition and nutrient distribution in humus profiles in some Mediterranean forests in southern Tuscany, Forest Ecology and Management 57: 99–114. and Veer, M. A. C. (1992), Soil organic-matter accumulation, litter decomposition and humus forms under Mediterranean-type forests in Southern Tuscany, Italy, Journal of Soil Science 43: 133–44. Mulligan, M., and Poesen, J. (2000), Spatial patterns of soil water balance on intensively cultivated hillslopes in a semi-arid environment: the impact of rock fragments and soil thickness, Hydrological Processes 14: 1811–28.

202

John Wainwright

van Wesemael, B., Verstraten, J. M., and Sevink, J. (1995), Pedogenesis by clay dissolution on acid, low-grade metamorphic rocks under Mediterranean forests in southern Tuscany (Italy), Catena 24: 105–25. Vavliakis, E. G., Haristos, D. A., and Balafoutis, C. (1990), Indirect influence of man-made factors on the dissolution rate of dolomitic marble in Thessaloniki area (northern Greece), Zeitschrift für Geomorphologie 34: 475–80. Wainwright, J. (1994), Anthropogenic factors in the degradation of semi-arid regions: a prehistoric case study in southern France, in A. C. Millington and K. Pye (eds.), Effects of Environmental Change on Drylands. John Wiley & Sons, Chichester, 285–304. (1996a ), Constitution et évolution des nappes archéologiques, in J. Gascó, L. Carozza, S. Fry, R. Fry, J.-D. Vigne, and J. Wainwright (eds.), Le Laouret et la Montagne d’Alaric à la Fin de l’Âge du Bronze, un hameau abandonné entre Floure et Monze (Aude). Centre d’Anthropologie, Toulouse, 251–60. (1996b), Infiltration, runoff and erosion characteristics of agricultural land in extreme storm events, SE France, Catena 26: 27–47. (1996c), Hillslope response to extreme storm events: the example of the Vaison-la-Romaine event, in M. G. Anderson and S. M. Brooks (eds.), Advances in Hillslope Processes. John Wiley & Sons, Chichester, ii. 997–1026. (1996d ), A comparison of the infiltration, runoff and erosion characteristics of two contrasting ‘badland’ areas in S. France, Zeitschrift für Geomorphologie Suppl. 106: 183–98. and Thornes, J. B. (2003), Environmental Issues in the Mediterranean: Processes and Perspectives from the Past and Present. Routledge, London. Mulligan, M., and Thornes, J. B. (1999), Plants and water in drylands, in A. J. Baird and R. L. Wilby (eds.), Ecohydrology. Routledge, London, 78–126. Parsons, A. J., Müller, E. N., Brazier, R. E., Powell, D. M., and Fenti, B. (2008), A transport-distance approach to scaling

erosion rates: 1. Background and model development. Earth Surface Processes and Landforms 33: 813–26. Whipple, K. X. and Meade, B. J. (2006), Orogen response to changes in climatic and tectonic forcing, Earth and Planetary Science Letters 243: 218–28. Willett, S. D., Slingerland, R., and Hovius, N. (2001), Uplift, shortening, and steady state topography in active mountain belts, American Journal of Science 301: 455–85. Wise, S. M., Thornes, J. B., and Gilman, A. (1982), How old are the badlands? A case study from southeast Spain, in R. B. Bryan and A. Yair (eds.), Badland Geomorphology and Piping. GeoBooks, Norwich, 259–77. Wondafrash, T. T., Sancho, I. M., Miguel, V. G., and Serrano, R. E. (2005), Relationship between soil color and temperature in the surface horizon of Mediterranean soils: a laboratory study, Soil Science 170: 495–503. Woodward, J. C. and Goldberg, P. (2001), The sedimentary records in Mediterranean rockshelters and caves: archives of environmental change, Geoarchaeology 16: 327–54. Macklin, M. G., and Lewin, J. (1994), Pedogenic weathering and relative-age dating of Quaternary alluvial sediments in the Pindus Mountains of northwest Greece, in D. A. Robinson and R. B. G. Williams (eds.), Rock Weathering and Landform Evolution. John Wiley & Sons, Chichester, 259–83. Yaalon, D. H. (1997), Soils in the Mediterranean region: what makes them different? Catena 28: 157–69. Yair, A. (1995), Short and long term effects of bioturbation on soil erosion, water resources and soil development in an arid environment, Geomorphology 13: 87–99. Yassoglou, N., Kosmas, C., and Moustakas, M. (1997), The red soils, their origin, properties, use and management in Greece, Catena 28: 261–78. Zalidis, G., Stamatiadis, S., Takavakoglou, V., Eskridge, K., and Misopolinos, N. (2002), Impacts of agricultural practices on soil and water quality in the Mediterranean region and proposed assessment methodology, Agriculture Ecosystems and Environment 88: 137–46.

This chapter should be cited as follows Wainwright, J. (2009), Weathering, soils and slope processes, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 169–202.

7

Vegetation and Ecosystem Dynamics Harriet Allen

Introduction Within the Mediterranean region a number of distinctive vegetation communities can be recognized, comprising some 25,000 species, of which about 50 per cent are endemic. Broadly defined, these originated with the establishment of a mediterranean-type climate about 3.2 million years ago, since when they have been subject to the vicissitudes of glacial–interglacial climate changes, plus the intensification of human impact during the last 10,000 years (Chapters 4 and 9). These communities are dynamic, responding to environmental changes at a variety of scales, both spatial and temporal. This chapter explores the characteristics of these communities and examines the relationships between ecosystem dynamics and biodiversity, and ecosystem response to disturbance. For example, each year fires burn out of control and are the subject of regular news stories during summer months. While fires may be economically devastating and lead to loss of life (Chapter 19), ecologically their incidence is an important dynamic component of Mediterranean ecosystems and may, indeed, be crucial to the successful propagation and spread of plants and communities regarded as typically Mediterranean. Associated animal populations generally recover quickly despite inevitable loss of life in some populations. Thus understanding the role of fire and other disturbance factors such as grazing is key to understanding Mediterranean vegetation communities and ecosystem dynamics. The chapter concludes with an evaluation of the likely response of vegetation communities to potential atmospheric and land use changes.

Vegetation Communities While a number of distinct vegetation communities have been identified, a common characteristic is an ability

to survive hot, dry summers and cool, wet winters, together with frequent disturbances. Many of the communities are dominated by shrubs, and Mediterranean evergreen sclerophyllous shrublands are recognized as one of the defined ecosystems of the world (di Castri 1981). Such shrublands are at the centre of a continuum of communities which vary along gradients of moisture availability, temperature, and nutrient availability, usually determined by substrate, and human activity (Figure 7.1). At the extreme ends of these gradients, but still Mediterranean, are sclerophyllous woodlands, coniferous and deciduous forests, savannas and grasslands grading into steppe and semi-desert shrublands, and heathlands. The relationship between vegetation and climate is recognized in bioclimatic classifications (Quézel 1985). Eight different life zones (Figure 7.2) are identified according to temperature and precipitation indices, varying with either latitude or altitude, and therefore recognizable in a north–south traverse across the Mediterranean basin or along an altitudinal transect. In each of these bioclimatic zones, some of the dominant species can be identified as characteristic bioindicators.

Evergreen Sclerophyllous Shrublands Many Mediterranean countries have their own names for evergreen shrubland communities, with the terms ‘maquis’ or ‘matorral’ widely used (Table 7.1). This can cause confusion in the literature, as the names are not defined precisely (Allen 2001), and may be compounded by classifications that depend not just on the life form of the species present (e.g. whether tree or shrub, or perennial or annual), but also on the height and proportion of ground cover. For example, maquis communities are generally regarded as comprising relatively tall-growing shrubs (Figure 7.3), while garrigue is a

204

Harriet Allen Savannas and grasslands Desert and semi-desert shrublands

Fertilization

Greater nutrient availability

Aridity Climatic/ environmental control Human intervention Desertification

Evergreen sclerophyllous shrublands

Coniferous and deciduous forests

Higher rainfall, lower winter temperature

Greater moisture availability

Low nutrient availability

Sclerophyllous woodlands

Heathlands

Fig. 7.1. Mediterranean vegetation communities and trajectories of change.

lower-growing community (Figure 7.4), but a continuum exists between the two. Mediterranean shrublands are dominated by sclerophyllous taxa. These have tough, leathery evergreen leaves, with a low surface-area to volume ratio, thick cuticles, and stomata sunken into grooves on the underside of the leaf. These characteristics are widely believed to be adaptations to water stress, as a means to control transpiration losses. Other selective forces might include response to fire, herbivory, and nutrient-poor substrates (Mooney and Dunn 1970). In addition to evergreens, deciduous shrubs and aromatic plants are common in shrubland communities. Some deciduous shrubs show seasonal dimorphism; that is, leaf morphology occurs in two forms as a response to intense summer drought. Rates of photosynthesis and respiration are higher in winter leaves due to higher concentrations of soluble sugars, chlorophyll, fat, nitrogen, and free and bound amino acids compared with summer leaves (Margaris 1977). Within a genus some species may be evergreen, while others are deciduous. For example, there are about ten deciduous Pistacia species, but only one evergreen, Pistacia lentiscus. Among the Mediterranean oaks, there are six evergreen species, including Quercus ilex, Q. suber, and Q. coccifera, about 35–40 deciduous, including Q. robur, Q. cerris, and Q. pyrenaica, and one semi-deciduous, Q. infectoria (Blondel and Aronson 1999). The Mediterranean flora contain some 49 per cent of the world’s aromatic taxa, which produce essential

oils. The majority belong to the mint (Lamiaceae), carrot (Apiaceae), and daisy (Asteraceae) families. The production of essential oils within leaves appears to peak in summer (Ross and Sombrero 1991), which suggests a regulatory factor associated with seasonality. This need not be drought, but might be a defence mechanism against herbivory, either by insects or grazing animals. The leaves of many aromatic plants are lacking in nitrogen, and as soils are often nitrogen- and phosphoruslimiting, leaf growth is slow. Production of long-lived leaves is therefore an advantage. As well as being a deterrent to herbivory, essential oils may be anti-microbial or anti-fungal, may attract pollinators by mimicking insect pheromones, and may cool leaves through volatilization of the oils, thus reducing transpiration. Maquis communities are the most widespread of the mediterranean ecosystems. They form a narrow coastal belt and penetrate inland according to climatic and topographic variability, being more extensive in the western than eastern Mediterranean (Figure 7.5). Their altitudinal and latitudinal limits appear to coincide with an average minimum temperature in the coolest month of 0◦ C. Summer maximum temperatures do not appear to be a limiting factor. Their extent usually coincides with mean annual precipitation levels of 350 to 1,500 mm, though in North Africa and the Middle East the limiting annual total is less than 200 mm. While nutrient availability may not be directly responsible for the distribution of maquis communities, they are absent from salty soils, and also from thin, highly eroded soils

Vegetation zonation

Vegetation communities

Cryo-Mediterranean

Alpine on rock, scree, and gravel

Alti-Mediterranean

Sub-alpine grasslands, herbaceous perennials, and dwar f junipers

Juniperus , Bromus, Festuca, Poa, Phleum

Oro-Mediterranean

Coniferous woodland

Pinus uncinata, Pinus mugo

Montane Mediterranean

Deciduous woodland

Fagus sylvatica; conifers (Pinus, Cedrus, Abies), Juniperus

Supra-Mediterranean

Meso-Mediterranean

Infra-Mediterranean (Western Morocco)

4,000

Saxifraga, Androsace, Aubretia Average minimum temperature of the coldest month (°C)

3,000

– 9°C

Quercus humilis, Quercus cerris, Deciduous oak forests Quercus frainetto, Quercus macedonia, and semi-deciduous in Ostrya carpinus, Carpinus orientalis, Nor th Africa and Spain Corylus sp ,Tilia sp., Fraxinus ornus, Acer sp.s Quercus ilex, Quercus calliprinos, Pinus halepensis, Pinus brutia. Where heavy Evergreen oak woodands anthropogenic activity: Quercus coccifera, and shrublands Calycotome villosa, Genista acanthoclada Olea europea, Ceratonia siliqua, Phillyrea Dense coastal woodand; media, Pistacia lentiscus, Laurus nobilis, sclerophyllous, evergreen; Quercus suber, Pinus pinaster, significant human impact Chamaerops humilis

2,000

–5°C –2°C +1°C

1,000 +3°C

Argania spinosa and Acacia gummifera

+5°C

+7°C

0 30

32

34

36 38 40 Latitude (°N)

Fig. 7.2. Bioclimatic life zones of the Mediterranean region (after Blondel and Aronson (1999) and Le Houérou (1990)) (see Chapters 5 and 23).

42

44

46

Vegetation and Ecosystem Dynamics

Thermo-Mediterranean

Altitude (metres)

Bioindicator taxa

205

206

Harriet Allen

TABLE 7.1. Terms used to describe Mediterranean sclerophyllous shrubland High matorral Israel Greece Italy Macchia alta France Maquis

Spain

Matorral denso, espinal

Middle matorral

Macchia bassa Garrigue (sometimes only with reference to calcareous soils) Matorral claro, jaral (on siliceous soils)

Low matorral Batha Phrygana Garriga Tomilar

Garrigue (or Landes on siliceous soils)

Source: After Tomaselli (1981a) and Margaris (1981).

at the climatic limits of their growth; a likely response to low soil moisture retention. Thus at higher elevations, they are found in association with calcareous soils rather than on poorly developed metamorphic-derived soils (Quézel 1981a, b). Typical evergreen or semi-evergreen trees and shrubs that have a circum-Mediterranean distribution (Figure 7.6) include olive (Olea europea), carob (Ceratonia siliqua), Pistacia, and some members of the Cistaceae family (Figure 7.7), e.g. Cistus salvifolius, and Ericaceae family, such as Erica arborea and Arbutus unedo. Many aromatic labiate species are also widespread, such as rosemary (Rosmarinus officinalis), lavender (Lavandula spp.), and thyme (Thymus spp.). Some taxa have more restricted distributions. Those more common in the western Mediterranean include species of Genista, Cytisus, and Ulex (all members of the pea family— Fabaceae), and several members of the Cistaceae and Ericaceae families. In the eastern Mediterranean, there are several labiate species of Phlomis, Satureja, and Salvia, and a greater number of species of shrubs such as Rhamnus and Daphne than are found in the western Mediterranean. The Judas tree (Cercis siliquastrum), though widely planted for ornament elsewhere, has its origin in the eastern Mediterranean. Four species of oak and two pines are widely found in sclerophyllous shrublands: holm oak (Quercus ilex), kermes oak (Q. coccifera), cork oak (Q. suber), and Aleppo pine (Pinus halepensis) in the western Mediterranean, and Q. calliprinos and Calabrian pine (P. brutia) in the eastern Mediterranean (Figure 7.6). These are examples of vicariant or disjunct distributions; that is, the distributions of two closely related but non-coexisting species. Conifers that grow in association with sclerophyllous shrublands include junipers (Juniperus phoenicia, J. oxycedrus). There is some indication that arborescent taxa are more common in the eastern than western Mediterranean (Quézel 1981a).

Sclerophyllous Woodlands Although evergreen oaks such as holm oak (Quercus ilex) may be present in the sclerophyllous shrubland communities, at the wetter and cooler margins of the Mediterranean they can form more or less continuous woodland cover (Terradas 1999). Q. ilex has a wide distribution, but is replaced in the eastern Mediterranean by Q. calliprinos. Holm oak forests are particularly extensive in Sicily, Italy, Corsica, France, Spain, and Morocco and, historically, have been of considerable economic importance, in particular as a source of charcoal (Terradas 1999). Pollen diagrams from Huelva in south-western Spain suggest that woodland management occurred as early as 6,000 cal. years BP (Harrison 1996), although in other areas of the Mediterranean basin such management might date from later periods.

Savanna Woodlands In addition to closed woodland, savanna-type landscapes of ‘trees without forests’ can be recognized across the Mediterranean—areas of trees scattered amongst other vegetation types. Grove and Rackham (2001) regard their origin on the Iberian Peninsula to be the result of management formalized over centuries to create dehesas (or montados in Portugal). Such a cultural origin is likely elsewhere but savanna woodlands also occur in the drier areas of the Mediterranean. Trees of savanna systems were traditionally managed by coppicing and pollarding and their continued existence was economically crucial: dehesas traditionally provided acorns and shelter for livestock, tannin, cork, and firewood, and pasture for sheep and cattle (Joffre 1992). Trees associated with Iberian dehesas include the evergreen oaks (Quercus ilex, Q. rotundifolia) and the cork oak (Q. suber) and deciduous oaks (Q. faginea and Q. pyrenaica). Conifers and deciduous trees are also characteristic of savanna woodlands. In the Alpujarra (Spain), Italy, Corsica, and western Crete, sweet chestnut (Castanea sativa) is important, as are beech (Fagus sylvatica) in Italy, Corsica, and northern Greece, and carob (Ceratonia siliqua) in southern Portugal and Crete (Grove and Rackham 2001). The emergence and longevity of savanna-type woodlands has resulted in high levels of biodiversity and they are often important conservation sites. But their continued existence, and that of associated fauna, is threatened by land use changes (Chapter 23).

Vegetation and Ecosystem Dynamics

207

Fig. 7.3. Invasion of sclerophyllous maquis vegetation into an old olive orchard, Crete (photo: Harriet Allen).

Coniferous Forests In mountainous areas and on coastal sandy soils, coniferous forests are widespread, although plantations also exist at lower elevations. The most widespread conifers are the pines, but other important taxa include cedars (Cedrus atlantica in the Atlas Mountains, C. brevifolia in Cyprus, and C. libani in Lebanon and Turkey) and, in high mountain regions, firs (Abies sp.). About ten different pine species grow around the Mediterranean, although only four are regarded as true Mediterranean species (Klaus 1989): Aleppo pine (Pinus halepensis), Calabrian pine (P. brutia), Canary Island pine

(P. canariensis), and stone or umbrella pine (P. pinea). Of these, Aleppo and Calabrian pines are another example of vicariant distributions (Figure 7.6). Genetically and ecologically similar, they coexist in two small areas of Greece, in south-eastern Anatolia, and in Lebanon and Israel, where they form natural hybrids (Barbero et al. 1998). Geological, archaeological, and historical evidence points to an eastern origin for both and climate is the likely determinant of their modern distribution (Biger and Liphschitz 1991). P. halepensis requires a more humid and warmer climate; it cannot survive temperatures of −10◦ C. It therefore spread west in response

208

Harriet Allen

Fig. 7.4. Typical garrigue vegetation in Crete. Note the lower-growing, rounded nature of the grazed shrubs (photo: Harriet Allen).

to its climatic needs and climatic change. By contrast, P. brutia can survive lower temperatures and therefore remained in its area of origin. However, its occurrence in Israel and the surrounding areas may be a recent phenomenon, given its absence from wood samples collected at Israeli archaeological sites during the late Holocene (Biger and Liphschitz 1991). Pine forests are estimated to cover about 5 per cent of the total Mediterranean, but comprise about 25 per cent of the forest cover in Anatolia and up to 75 per cent in North Africa (Barbero et al. 1998). Their extent has changed considerably in recent decades. In North Africa,

decline has followed clearance for cultivation, felling of timber for construction and charcoal production, overgrazing, and frequent fires. By contrast, in the European Mediterranean, pine forest has increased in area as a result of land use changes, which include afforestation, and the ability of pine to invade abandoned or burned land (Barbero et al. 1998).

Grasslands, Steppe, and Semi-desert Shrublands Around the Mediterranean basin, grasslands tend to occur where mean annual precipitation is less than

Vegetation and Ecosystem Dynamics

0

209

500 km

Fig. 7.5. Extent of maquis communities across the Mediterranean region (after Quézel 1981a).

about 300 mm (Blondel and Aronson 1999). In addition, perennial grasses form a significant component of ‘alti-mediterranean’ or subalpine communities (Figure 7.2). Some of the most extensive grasslands are found in North Africa and Spain, such as those dominated by the perennial bunch grass, Stipa tenacissima (alfa grass). This is a slow-growing drought- and high-temperature endurer, which mainly regenerates vegetatively and grows as tussocks. Between the tussocks, annual grasses invade. These persist until the Stipa tussocks have encroached into the inter-tussock spaces, while the older tussocks become senescent. The dying tussocks then form new inter-tussock areas, to be invaded by annual grasses. This two-dimensional regeneration pattern is believed to confer stability on the Stipadominated grasslands; the presence of annuals, albeit at different densities and biomass, ensures that few areas of bare ground exist for long, which reduces the potential for soil erosion (Clark et al. 1998; Chapter 6). In especially arid areas, such as parts of Spain and North Africa, Stipa and Brachypodium grasses grow in association with shrubs such as wormwood (Artemisia). In more arid areas, grassland species grow in association with other arid taxa, such as Artemisia, Pistacia, and Ephedra. These are identified as being Irano-Turanian

in origin; that is, their centres of origin are the semiarid steppes of central Asia, where summers are exceptionally hot and winters very cold and dry. Today, the Irano-Turanian phytogeographical region extends as far west as Jordan and the non-coastal parts of Lebanon, Israel, and Syria. Formerly, during the glacial stages of the Quaternary it spread further west across the Mediterranean region. Thus the steppe communities, identified by pollen analysis, at the height of the global Last Glacial Maximum in northern Europe, about 18,000– 20,000 years ago, were indicative of extensive Mediterranean aridity. A large number of Irano-Turanian elements are recognized within the contemporary Mediterranean flora (Blondel and Aronson 1999) and these are believed to represent the successful ‘invasions’ of xeric plants into more humid, mediterraneantype communities following human activity; disturbed ground is susceptible to invasion by weedy taxa more representative of the drier habitats of the Mediterranean. This is referred to by Blondel and Aronson as Zohary’s ‘law’ based on the ‘expansion drive’ of desert plants into the Mediterranean region of the Near East (Zohary 1962). This expansion of grasses and associated taxa underlies the recognition of Mediterranean grasslands and steppe as anthropogenic (Figure 7.8), being

210

Harriet Allen Olea europea subsp. oleaster

0

Arbutus unedo

1000 km

Cistus salvifolius

Lavandula stoechas

Quercus ilex Quercus calliprinos

Pinus halepensis Pinus brutia

Quercus suber

Cercis siliquastrum

Fig. 7.6. Distribution maps for circum-Mediterranean taxa, Olea europea subsp. oleaster (wild olive), Arbutus unedo (strawberry tree), Cistus salvifolius (sage-leaved cistus), Lavandula stoechas (lavender), and vicariant taxa, Quercus ilex and Q. calliprinos (holm oaks), Pinus halepensis (Aleppo pine) and P. brutia (Calabrian pine), Quercus suber (cork oak), and Cercis siliquastrum (Judas tree). Based on Daget (1980), Quézel (1985), and Biger and Liphschitz (1991).

Vegetation and Ecosystem Dynamics

211

Fig. 7.7. Cistus ladanifer-dominated scrub vegetation of the Algarve, Portugal (photo: Harriet Allen).

degraded maquis communities. In areas of hyper-aridity, where saline soils form due to high water tables, salttolerant taxa grow, many of which are members of the Chenopodiaceae (goosefoot) family.

Heathlands Heathlands may occur where nutrient availability is low, for example on poor siliceous soils. Although Mediterranean heathlands received only limited mention in the review of global heathland ecology (Specht 1979), they are recognized in phytosociological surveys by their typical vegetation associations, especially on the sandy soils of France, known as landes. Typical genera are Erica, Cistus, Quercus, and Genista. Sandy soils are generally acidic, which limits the availability of phosphorus, compared with neutral to alkaline soils. When such nutrient-poor soils are fertilized,

productivity increases (di Castri 1981) which demonstrates the nutrient-limiting characteristics of Mediterranean heathlands. The extent to which Mediterranean heathlands are natural or a result of human impact is debatable. Pollen analysis indicates that they can occur spontaneously or have a long history. Low-growing shrub formations of Pistacia, Phillyrea, Juniperus, and Thymelaea have been identified in communities that lacked trees such as oak (Quercus), fir (Abies), or beech (Fagus). This suggests the heath taxa were not merely representative of woodland understorey communities, but existed as communities in their own right (Blondel and Aronson 1999). Mediterranean heathlands generally occur as isolated islands surrounded by more calcareous substrates. Floristic surveys of the heathlands of the Strait of Gibraltar region of Spain show them to have high levels of species endemism, though low species richness, due to

212

Harriet Allen Primary maquis/ sclerophyllous evergreen shrubs

Regeneration Reasonable exploitation

Secondary maquis

Degeneration

Excessive exploitation

Garrigue

Grazing

Fire

Cultivation

Steppe

Abandonment

Pasture/grassland Cultivation

Degeneration Regeneration

Regeneration

Abandonment

Orchards/crops

Fig. 7.8. Theoretical degradation and regeneration sequences for primary maquis or sclerophyllous evergreen shrub communities.

the nutrient-poor status of the acid, sandy soils (Ojeda et al. 1995). This could make them special targets for conservation.

Sub-alpine, Alpine, and Cliff Communities As a highly mountainous region, sub-alpine and alpine communities form an important component of Mediterranean flora. In the sub-alpine zone, a common characteristic is that of ‘hedgehog’ plants, a term used to describe their spiny, dwarf form (Figure 7.9). Many of these are members of the Fabaceae (pea) family, such as Astragalus angustifolius (a milk vetch) and Genista acanthoclada, both found in Crete. The ‘hedgehog’ zone is well developed in the Pyrenees, Sierra Nevada mountains of Spain, in the Moroccan Atlas mountains, and in Greece, Crete, and the Taurus mountains of Turkey. Growing alongside these plants are often rich swards of perennial bunch grasses. Where rainfall is heavier, mountain grasslands form important pasture land for summer grazing; transhumance remains an important part of livestock management in the Iberian Peninsula (Ruíz and Ruíz 1986) and in Greece and Crete (Rackham and Moody 1996). Within the sub-alpine zone there may also be shrub forms of trees such as Juniperus communis subsp. nana. Many of these communities have been influenced by fire and grazing.

True alpine communities are found at the highest elevations, for example above 2,600 m in the Pyrenees and Sierra Nevada (Polunin and Smythies 1973) and above 2,200 m on Crete (Sfikas 2002). The typical growth habit of alpine plants is that of prostrate, creeping, rosette, or cushion form sheltering amongst areas of bare rock, boulders, gravel, and scree. Typical alpine genera include Saxifraga, Androsace, and Aubretia. In addition, geophytes (bulbs) are a relatively common element of alpine and subalpine communities. A common characteristic of Mediterranean alpine communities is the high numbers of endemic species—a result of the geographical and reproductive isolation of these high elevation areas—which makes them of great interest to botanists. The Pyrenees have over 180 endemic alpine species (Polunin and Smythies 1973) and a similar number are found on Crete (Rackham and Moody 1996). Cutting through many mountainous areas are deep ravines and gorges, for example the Samaria and Imbros gorges of Crete, and the Vikos Gorge in the Pindus Mountains of Greece. Steep cliffs, with steps and crevices, create strong microclimatic gradients, which, together with factors such as aspect, lead to a variety of habitats and assemblages of plants, many of which are out of reach from browsing and grazing animals. These sites are often rich in endemic species and have frequently acted as refugia for taxa which have undergone migrations and population recessions through glacial– interglacial cycles (Chapter 4). For example, Tertiary

Vegetation and Ecosystem Dynamics

213

Fig. 7.9. The typical ‘hedgehog’ shape of the alpine, Euphorbia acanthothamnos, growing among rocky scree of the Psilorites Mountains, Crete (photo: Harriet Allen).

relicts such as Viola cazorlensis, Pinguicula vallisneriifolia, and Ptilotrichum reverchonii are found in the gorges of the Sierra de Cazorla, southern Spain (Polunin and Smythies 1973). The inaccessibility of the high cliff faces of the gorges, which may extend from sea-level up to the sub-montane and montane zones, means that they have remained little touched by human activity and so are important conservation sites.

Wetlands Wetland vegetation communities, of international importance for indigenous wildlife and migrating wildfowl, occur around the Mediterranean. As they differ according to duration of flooding, depth of water column, and salinity, there is a continuum between saline coastal marshes and drier, more freshwater marshland. Notable wetland communities of the

Mediterranean region include the marismas (annually flooded marshlands) of the Coto Doñana in Spain, the Camargue of France, Amvrakikós of Greece, Lakes Bardawil and Burullus in Egypt, and Garet el Ichkeul and Sebkhet Kelbia in Tunisia (Chapter 9). The future of some of the larger wetlands is assured because of their importance to conservation, but many of the smaller ones are transient, both in terms of their ecological status as a response to changing sea levels and because of catchment land use changes which alter their hydrological balance.

‘Semi-natural’ Landscapes Agriculture spread across the Mediterranean from the Near East during early to mid-Holocene times. Consequently few vegetation communities can be regarded as untouched and some have developed alongside

214

Harriet Allen

cultivation practices, leading to ‘semi-natural’ landscapes. Examples include olive groves and terraced landscapes often associated with olives and orchard trees. In Crete traditionally managed olive groves have formed part of the island’s landscape ecology for thousands of years (Allen et al. 2006). Pollen evidence confirms cultivation as early as 5,700 (uncalibrated) years BP, before the beginning of the Early Minoan I period and earlier than on either the Greek or Turkish mainlands (Bottema and Sarpaki 2003). Such groves often support a rich ground flora, which some researchers believe resembles that of ‘natural’ Mediterranean ecosystems (Loumou and Giourga 2003). The ground flora is characterized by annuals and geophytes (bulbs). The annuals survive because regular shallow tilling and disturbance of the soil promotes germination of seeds; among the annuals are a number of species regarded as rare weeds (Phitos et al. 1995). Geophytes survive because they are buried beneath the depth of tillage; for example, species of the Serapias orchid. The diversity of flora provides important habitats for a variety of fauna, such as mammals, birds, and insects (Loumou and Giourga 2003). Terraced landscapes (Figure 7.10) developed as an inevitable response to complex topography, and terraced agriculture is now integral to large areas of the Mediterranean region (Grove and Rackham 2001). Some terracing is ancient and based on the construction and maintenance of retaining walls; other terracing is more recent and relies more on the use of bulldozers. A variety of cultivation can be practised and mixed on terraces, including olives, vines, orchard trees, and arable crops. As with olive groves, some of these can develop speciesrich ‘semi-natural’ vegetation communities. However, maintenance of terraces is often labour intensive and much of it is now falling into disrepair, as terraces are abandoned—a consequence of twentieth-century agricultural intensification on more easily cultivable land (Allen et al. 2006). Abandonment has led, in places, to invasion by perennial shrub and bush species, creating maquis communities (Figure 7.3). The consequence of collapsing terraces has also fuelled a debate about their role in soil erosion, as illustrated by an example from the Alpujarra, in Andalucia. Historically, four periods of neglect of terraces have been identified as responsible for increased soil erosion (McNeill 1992). One of these came after the expulsion of the Moors from Spain in the fifteenth century AD. The Moors had established a system of well-tended terraced and irrigated agriculture, which subsequently declined leading to erosion of topsoil and its deposition in the deltas of the Mobril and Adra rivers. More recently, population decline in the region and abandonment of terraces, following

agricultural deintensification since the Second World War, are believed to have led to increased soil erosion (McNeill 1992). However, there is some uncertainty about rates of erosion and sediment yield in the Alpujarra, as these are measured as having a high degree of spatial variability (Douglas et al. 1994). This makes it difficult to extrapolate them and generalize with respect to the role of terracing. Indeed there is a largely positive view that terrace abandonment may reduce runoff and erosion of topsoil as maquis vegetation increases in extent (Thornes 1998). The diversity of Mediterranean biomes and vegetation communities means that there are numerous ecotones, or transitional habitats formed by the overlapping areas of two adjacent communities. These may be small in area, but this adds to the importance of ecotones and serves to increase the spatial heterogeneity of Mediterranean communities—there are large differences in animal and vegetation communities across short distances. The locations of ecotones are determined by factors such as climate, soils, and disturbance regimes, for example grazing and fire. Such variability adds considerably to landscape diversity and is a reflection of the ecosystem dynamics occurring within different communities.

Plant Biodiversity and Ecosystem Dynamics Empirical research, in a variety of world biomes, suggests that there are strong reasons to assume that biodiversity has an influence on ecosystem dynamics, although trying to establish the relationships between the two is difficult, especially as some of these relationships might be multi-variable in cause and scale-dependent. Assessing the possible influence of biodiversity on ecosystem stability helps to inform the debate about the fragility/resilience of Mediterranean ecosystems. The consequences of this need to be assessed as models of global biome change predict that mediterraneantype ecosystems will probably experience a large loss of biodiversity through the twenty-first century (Sala et al. 2000; and Chapter 23). It may be that Mediterranean ecosystems are becoming (even) more vulnerable to change. However, this has to be examined in the context of Mediterranean landscapes and ecosystem stability, which have traditionally been regarded as vulnerable but are now coming to be recognized as relatively robust. Grove and Rackham (2001) illustrate this debate in a discussion of the question of ‘ruined landscapes’: millennia of human activity have supposedly led to degradation of Mediterranean landscapes,

Vegetation and Ecosystem Dynamics

215

Fig. 7.10. Olive terraces on Crete (photo: Harriet Allen).

when in fact many plant communities and landscapes have developed under different patterns of human occupancy and variable climate, to the extent that they have become quite resilient. It is likely that this resilience is promoted by the region’s rich floristic diversity. Hence there is a need to examine why the region is so floristically rich (see also Chapter 23) and to assess the research on the influence of biodiversity on ecosystem processes and stability in the Mediterranean context. There are an estimated 25,000 species of flowering plants and ferns in Mediterranean Europe, of which about 50 per cent are endemic, in an area of about 2.3 million km2 , compared with just 6,000 species in the 9 million km2 of non-Mediterranean Europe (Quézel 1985). Nested within the region are nine diversity hotspots identified on the basis of numbers of endemic plants that appear to be threatened. These are the High and Middle Atlas Mountains, The Betic-Rif complex,

the Maritime and Ligurian Alps, the Tyrrhenian Islands, southern and central Greece, Crete, Anatolia and Cyprus, Syria–Lebanon–Israel, and Mediterranean Cyrenaica in Libya (Médail and Quézel 1997). No single explanatory factor for species richness can be advanced, as the environmental variables predominantly responsible apply at different spatial and temporal scales (see for example Willis and Whittaker 2002). For the Mediterranean region relevant factors include geological and palaeogeographical factors (such as the Alpine mountain building episode leading to reproductive isolation of populations), the consequences of environmental changes (glacial–interglacial cycles), the importance of ecological stress and competition between species, the diversity of habitats (especially relating to topography, microtopography, and microclimate), and the role of disturbance (for example, fire).

216

Harriet Allen

Geological and palaeogeographical factors are normally cited to explain continental- to regional-scale diversity patterns. For example, some taxa that are typically regarded as Mediterranean, in fact evolved prior to the establishment of a mediterranean-type climate: pollen grains from olive (Olea), Phillyrea, Cistus, and Pistacia, are present in off-shore cores from the western Mediterranean Sea in deposits older than 3.2 million years (Suc 1984; Chapter 4). Some mediterranean-type taxa also appear to have their origins in non-Mediterranean regions, for example asparagus (Asparagus), carob (Ceratonia), and oleander (Nerium), which are believed to have spread from Africa at a date earlier than the Oligocene and Miocene (Raven 1973). The location of the Mediterranean region, at the crossroads of Europe, Asia, and Africa, has facilitated the spread of flora from other regions. Other historical explanations for an enriched regional biota are based on the response of taxa to environmental changes, such as those associated with the glacial–interglacial cycles of the Quaternary. The forests of southern Europe acted as refugia for tree species from northern Europe during glacial periods. Around Lake Ioannina in north-western Greece, for example, deciduous forest communities included oak (Quercus), elm (Ulmus), hornbeam (Carpinus betulus), and oriental hornbeam (C. orientalis) which had disappeared from the forests of northern Europe (Tzedakis 1993). Greater diversity in the Mediterranean applies at the genetic as well as at the specific level. Genetic differentiation of tree populations shows higher maternally inherited chloroplast DNA among the refuge populations of southern Europe (Thompson 1999; Hewitt 2000). However, greater diversity is not necessarily found in all measures of genetic diversity in refugial populations (Comps et al. 2001), which complicates the interpretation of molecular evidence for refugia. The varied topography and tectonic history of the Mediterranean region, which have created isolated populations, also help to explain plant species richness and patterns of endemism at regional to subregional and local scales, particularly in mountainous areas. Geographical isolation can lead to reproductive isolation allowing allopatric speciation and the evolution of endemics to occur. On Crete 26.6 per cent of the plant species found at elevations higher than 1,500 m in the Lefka Ori, or White Mountains, are endemic to the island; nested within this, 10 per cent are endemic to individual mountain massifs within the range (Vogiatzakis et al. 2003). The considerable topographic variability, especially in mountainous areas,

means that there are steep ecological gradients, providing a diversity of physical habitats for plants and animals. This promotes species richness at the local to landscape scales. Nested within this are local scale, short-term processes, resulting from factors such as fire and grazing, which are discussed in more detail below. From this brief review of reasons for biodiversity patterns it should be apparent that variables accounting for species richness differ with scale, both temporal and spatial, and that this needs to be taken into consideration when trying to examine the potential role of biodiversity in ecosystem dynamics.

Diversity and Ecosystem Functioning Identifying the components of biodiversity that are related to ecosystem functioning is currently the subject of much debate (see e.g. Schulze and Mooney 1994). In part this is because of the difficulty of defining the term ‘biodiversity’ and in part because of the difficulty in establishing what to measure. One of the commonest usages of the term is to mean ‘species richness’, and it is generally believed that greater plant species richness leads to greater productivity in plant communities, which leads to greater nutrient retention, and greater ecosystem stability (see e.g. Tilman 2000). As yet there have been few empirical studies across a range of ecosystems to test the relationships between diversity and stability. However, some results are being derived from grassland manipulation experiments in Europe, including the Mediterranean region. These suggest that species richness contributes to net primary productivity, but that so too does the type of species grown, i.e. whether perennial or annual, or whether the grassland herbaceous seed mix includes nitrogen-fixing taxa. The results of grassland manipulation experiments revealed a positive relationship between species richness and above-ground primary productivity (biomass) when measured in eight field trials across Europe. A reduction in the number of species grown in a plot resulted in a log-linear reduction in biomass (Hector et al. 1999), although the results were less conclusive for the Mediterranean field plots in Portugal and Greece compared with those in northern Europe. Experiments manipulating the type of grass grown at the sites on Lesvos, Greece, attempted to assess the role of functional diversity through inclusion or exclusion of perennial and annual grasses (Troumbis et al. 2000). Plots containing the perennial grass, Phalaris coerulescens, recorded the greatest primary productivity, suggesting that it may

Vegetation and Ecosystem Dynamics

be the presence of particular types of species that could be a determinant of productivity performance, rather than species richness per se. However, in reality many Mediterranean grasslands represent secondary successional communities, which are dominated by annual grasses, especially at the earliest stages if herbivory and grazing occur. Perennials may arrive at later stages and can become dominant through their more efficient use of water and nutrients. At this point their presence may mask or overwhelm subtle relationships between the diversity of more minor species and primary productivity. In fact, annuals and perennials can coexist in mosaic patches, as described in the example of Stipa-dominated grasslands (in the description of grassland communities above) where their coexistence is believed to confer stability (Clark et al. 1998). Manipulation experiments have also examined the influence of other measures of functional diversity, such as the presence of nitrogen-fixing legumes. CrossEuropean grassland experiments have shown that the presence of leguminous plants increases nitrogen accumulation, though the results were weakest for Greek and Portuguese sites compared with others (Spehn et al. 2000), possibly due to low availability of phosphorus which is necessary for adequate root nodule development. Where phosphorus and other nutrients are freely available, the presence of legumes generally increases nitrogen availability to non-legumes, although this depends on the legumes present as they differ in their efficiency of nitrogen-fixation. At some of the sites, legumes performed a key productivity role, thus supporting their identification as keystone species. In Mediterranean ecosystems, they commonly occur with grasses, with mutualistic associations developing. Together they are especially abundant in the early and middle successional stages (Blondel and Aronson 1999). Many leguminous taxa are members of the Fabaceae (pea) family, which contributes an important component to Mediterranean floral species richness. Globally, Fabaceae contains some 8,000 taxa in over 600 genera; a large proportion, though not all, are nitrogenfixing herbaceous and shrubby legumes, both annual and perennial. These are often present in early postdisturbance communities and their contribution to four regional floras of the Mediterranean has been assessed (Blondel and Aronson 1995). In the more arid regions, such as Tunisia and Israel, they accounted for 12.5 and 14.9 per cent of the regions’ Mediterranean elements respectively, compared with 9 per cent in both of the more humid regions of the Hérault (France) and Catalonia (Spain). The herbaceous perennial legumes were especially abundant.

217

There is a danger in extrapolating from field experiments in one or a few types of ecosystem to other ecosystems, but the results begin to support the idea that more diverse ecosystems are more productive and more stable. However, the mechanisms by which species coexist and maintain diversity are still the subject of much debate, and despite the abundance of hypotheses there has been no clear demonstration of actual maintenance processes in species-rich ecosystems (Tilman 2000). In the context of the Mediterranean region, a possible hypothesis as to how this operates relates to the existence of spatially heterogeneous habitats in which more species can coexist because of niche differentiation, because species differ in their resource use. Habitat differences, due to steep environmental gradients in factors such as microclimate, relief, soil moisture availability, etc., mean that each species performs best in only a portion of sites. As habitat or niche heterogeneity increases further, so too does species diversity leading to greater efficiency of resource capture and use; diversity increases the likelihood that species that are better able to exploit conditions are present, and the actions of different species in different niches complement one another. From this it might follow that lower species diversity could reduce plant productivity through declining niche complementarity and less efficient use of resources. This suggests that the variety of habitat and landscape elements in the Mediterranean region promotes species richness, which in turn promotes ecosystem stability (see Zamora et al. 2007). Of course, a range of interrelated factors influences ecosystem functioning. In addition to biotic composition, i.e. vegetation communities and their associated animals, there are abiotic factors, for example climate and soil type, and a variety of disturbance regimes. Historically, Mediterranean habitats and landscapes have been long subjected to disturbances, such as fire, grazing, and a range of human activities, which should therefore be regarded as positive factors in maintaining stability. This is not to argue that human activity does not result in landscape degradation, more that over the millennia of the Holocene it has done much to promote it (Grove and Rackham 2001). Modern trends, however, may be polarizing towards more intensive activity in some places, but with abandonment of landscapes in others. Increasingly global changes, such as climate change, elevated concentrations of atmospheric carbon dioxide, nitrogen deposition, and, potentially, changing levels of ozone within both the troposphere and stratosphere are also important. The consequences of all these changes might be increasingly homogeneous landscapes, with reduced species richness and potentially, therefore, reduced ecosystem stability.

218

Harriet Allen

The Role of Disturbance in Mediterranean Ecosystems Some of the highest levels of species richness anywhere in the world are found in moderately grazed shrublands and woodlands of the eastern Mediterranean. This applies at the level of alpha diversity, that is the number of species found within local assemblages or communities (Naveh and Whittaker 1979), and is very likely to be associated with repeated disturbance events such as fire, grazing, and felling, and with the incidence of drought in a semi-arid climate. This has led to recognition of perturbation-dependent ecosystems: maintenance of ecosystems requires perturbations, such as fire and grazing, to survive. Post-disturbance vegetation dynamics return communities to their pre-disturbance state, maintaining diversity (Naveh 1994, 1995). If disturbance ceases, then diversity declines with a reduction in the efficiency of ecosystem functioning. From the ideas outlined in the previous section, this might lead to a reduction in productivity as well as a reduction in the resilience of such ecosystems. The role of perturbation in ecosystem functioning is believed to operate at small scales, such as the disturbances caused by ant hills, herbivory by snails and insects, holes dug by small mammals, as well as at the larger communitylevel scale (Blondel and Aronson 1999). Thus perturbation appears to be a principal factor explaining the spatial heterogeneity of Mediterranean landscapes (Chapters 5 and 23).

Fire as a Disturbance Factor The identification of perturbation-dependent ecosystems has led to a revision in thinking about the role of fire (see also Chapter 19). Fires burning out-of-control in the Mediterranean region are regular news items during summer months and may, of course, be economically devastating and lead to loss of life. Prior to the

1980s it was estimated that as much as 10 per cent of all forests and shrublands burnt annually (di Castri 1981). These figures may now be higher as fire frequency appears to be increasing. Since the 1960s, the number of fires and the surface-area burnt in the European Mediterranean have increased exponentially (Pausas and Vallejo 1999). However, the inter-annual and inter-regional variability of fire occurrence is high. Fire is common in all the world’s mediterraneanclimate regions, though in the Mediterranean region is less likely to be natural, for example following lightning strikes. Instead, fire has been associated with people since at least the Neolithic (Naveh 1975) and possibly for as long as 300,000 years (Trabaud et al. 1993). In consequence, mediterranean-type ecosystems are fireadapted with species surviving, regenerating, and reproducing after fires (Table 7.2 and Figure 7.11). Although it is difficult to be certain as to whether the adaptations are to fire itself or to other disturbances, certain characteristics may be more pronounced as a result of the frequency or intensity of fire (Trabaud1987). Flammability is a measure of the ability of a fuel type to ignite and sustain a fire. Laboratory tests on the leaves of twenty-four dominant Mediterranean taxa have been used to categorize flammability (Table 7.3). Leaves are the most flammable part of a plant during the initiation and propagation of a fire due to their high surface-area to volume ratio and essential oil content, but the role of moisture is also important. The fresh foliage of Mediterranean shrubs generally becomes ignitable when moisture content drops below 75 per cent, and the potential for ignition is significantly related to moisture content of both live and dead fuels in grasses, shrubs, and forests. Moisture acts as a heat sink, as heat is used for the evaporation of moisture content. Moisture also dilutes volatiles and excludes oxygen from the area of combustion. The effect of moisture is determined, to an extent, by leaf structure. Least flammable taxa are those with adaptations to reduce water loss, or those with a parenchyma rich in cellular water in which more heat is required

TABLE 7.2. Fire-adapted strategies of some selected Mediterranean taxa Vegetative resprouters Obligate Quercus coccifera Pistacia lentiscus Ceratonia siliqua

Facultative Cistus salvifolius Salvia triloba Thymus capitatus Erica arborea Arbutus unedo Arbutus andrachne

Source: After Naveh (1974).

Stimulation of flowering

Stimulation of seed germination

Many geophytes—members of the following families: Iridaceae Lilliaceae Amaryllidaceae

Pinus halepensis Cistus monspeliensis Cistus albidus Rosmarinus officinalis Lavandula stoechas

Increased seed liberation and dispersal Pinus halepensis Pinus pinaster Pinus brutia

Vegetation and Ecosystem Dynamics

219

Fig. 7.11. Basal regrowth of Arbutus unedo in the spring of 2004, following fire in the summer of 2003, Monchique, southern Portugal (photo: Harriet Allen).

to evaporate the fuel moisture. Taxa with high levels of essential oils are amongst the most volatile, including Eucalyptus camaldulensis. Eucalyptus is a widely planted, introduced genus of tree that exacerbates fire risk. Compared with foliage, branches of typical Mediterranean taxa have higher particle density, a lower surface-area to volume ratio, lower total and silica-free ash content, and a lower heat content. As a result branches are less flammable (Dimitrakopoulos 2001). The ‘literature on fire is replete with generalizations regarding fire effects and vegetation response. Like the politician who searches the scriptures for phrases to match an ideology, the assiduous shrubland ecologist can find data to support any plausible and some implausible generalizations’ (Christensen 1985: 86). This situation is not the result of poor research but is more likely due to the high variability of shrubland fire regimes. Nevertheless, in mediterranean-type ecosys-

tems, studies of the vegetation dynamics of post-fire communities show that they rapidly return to a state similar to that of the pre-fire community in terms of species composition (Trabaud 1994). There is a large abundance of herbaceous taxa in the first few years after a fire, and many of these go on to become dominant members as the community matures, although gradually the structure of the community becomes more complex, with division into numerous vegetation layers. Fire has an influence on soil processes as well as vegetation dynamics and interactions occur between the soils and vegetation. Christensen (1994) summarized some of the effects of fire on Mediterranean soils as follows. Loss of vegetation and soil surface litter increases the intensity of raindrop impact and the amount of precipitation that reaches the ground. Consequently, there may be greater soil erosion immediately after a fire. This can lead to loss of sediment-bound nutrients as

220

Harriet Allen

TABLE 7.3. Flammability of selected Mediterranean plants based on laboratory tests of leaf ignition of Cretan species Species

Common name

Leaf characteristics

Less flammable Calicotome villosa Sarcopoterium spinosum

Spiny broom Thorny burnet

Juniperus oxycedrus Tamarix smyrnensis Castanea sativa Nerium oleander Platanus orientalis

Prickly juniper Tamarisk Sweet chestnut Oleander Plane

Thorns and spines which have high lignin content and low water permeability Extremely high silica-ash content

Moderately flammable Quercus coccifera Cistus salvifolius Cistus creticus Phlomis fruticosa Ceratonia siliqua Pistacia lentiscus

Kermes oak Sage-leaved Cistus Jerusalem sage Carob tree Lentisk

Hard leathery leaves with waxy or hairy epidermis which prevents water loss from evapotranspiration

Flammable Pinus brutia Pinus halepensis Quercus ilex Quercus pubescens Cupressus sempervirens Olea europea Erica arborea Arbutus unedo Pistacia terebinthus

Calabrian pine Aleppo pine Holm oak White oak Funeral cypress Olive Tree heath Strawberry tree Turpentine tree

Foliage has high surface area to volume ratio which facilitates water loss and heat absorption

Bay/laurel Eucalyptus

Rich in essential oils

Extremely flammable Laurus nobilis Eucalyptus camaldulensis

soil communities through erosion, unless they are too rapidly and intensively grazed by sheep and goats.

Grazing as a Disturbance Factor

Source: After Dimitrakopoulos and Papaioannou (2001).

topsoil is removed and to leaching of ions as rainfall infiltrates and percolates through the soils. Some nutrients, such as those containing nitrogen and sulphur, are lost through oxidation of organic compounds to gaseous form. Solid compounds may also vaporize, depending on the temperature of the fire, and nutrients may be lost further as ash blows away. However, not all nutrients are removed and those remaining contribute resources for the first plant colonizers. Many of these are nitrogenfixing legumes. Thus there is often a burst of flowering activity soon after fires, which aids the rapid regeneration of fire-prone vegetation communities. There may also be greater microbial activity and decomposition in the soils resulting in higher rates of mineralization and transformations such as nitrification, unless the fires have been so hot as to eliminate soil microbes. The occurrence of fire maintains a mosaic of communities, provided that the frequency is not too great. Fires do not necessarily have adverse effects on vegetation and

As characteristic Mediterranean disturbance factors, grazing and browsing are frequently considered alongside the effects of fire. The actions of sheep and goats (Figure 7.12) have traditionally been viewed negatively, being responsible for vegetation and soil degradation, even desertification (Margaris et al. 1996). However, this view should not be accepted uncritically. Like fire, grazing has been integral to Mediterranean communities since sheep and goats were domesticated some 10,000 years ago (Legge 1996; Uerpmann 1996). Moderate levels of grazing promote species diversity (Bergmeier 1998) and protect some open and semi-open communities from invasion by trees and woody species. Prior to 1948, grazing and woodcutting were widespread in the foothills of the Judaean Mountains of central Israel. This was followed by a period of effective control of wildfires and only limited grazing (Perevolotsky and Haimov 1992). By the mid-1980s, dense evergreen woodland had developed, composed mainly of dwarf shrubs (commonly Cistus species), moderate-sized shrubs (such as Pistacia lentiscus), and tall shrubs (dominated by Quercus calliprinos and Phillyrea latifolia up to 3–4 m in height). A consequence of the encroachment of woody species was a reduction of overall species diversity per unit area as less light was able to reach the ground flora, as well as increased accumulation of potential fuel for fires. Moderate grazing and fire are essential for maintaining species richness and landscape diversity, with vegetation and associated animal communities being the result of closely interwoven natural and cultural processes (Chapters 5 and 23).

Ecosystem Response to Introduced and Invasive Species Since the first trading routes were established between the Mediterranean and other parts of the world, exchanges of plants and animals have taken place, sometimes deliberately, other times accidentally (Groves and di Castri 1991). Many introduced taxa have become naturalized, or have economic value as crops or domesticated animals, or have aesthetic appeal. Some, however, have become pests. Mediterranean ecosystems, in general, appear to be prone to invasive taxa but the Mediterranean basin is less affected than other regions such as the Cape Floristic Kingdom (Cowling

Fig. 7.12. A goat browsing on Quercus coccifera, Psilorites, Crete (photo: Harriet Allen).

222

Harriet Allen

1992). Nevertheless, there are invasive species that pose a variety of ecosystem problems. There are some suggestions that abandoned fields may be more prone to invasive taxa (Mooney et al. 2001) particularly by ‘weedy’ species that take advantage of the disturbed and often fertile soils. Around the Mediterranean, two successful weeds are corncockle (Agrostemma githago) and knapweed (Centaurea cyanus). They spread into the Mediterranean basin from the Near East as agriculture spread—further examples of Zohary’s law on the expansion of Irano-Turanian flora (see above)—making the most of available nutrients, lack of competition, and an opportunity to set seed before being destroyed by ploughing or harvesting. These species spread in the early Holocene, and arguably should no longer be regarded as invasive, but more recent invaders are also found in abandoned fields, such as Conyza sumatrensis and C. canadensis, both erygerons, members of the Asteraceae (daisy) family. Both originate from the New World and have become widespread throughout Europe in the last 150 years. In southern France, C. sumatrensis is established and persists on old fields, while C. canadensis is restricted to recently disturbed sites (Thébaud et al. 1996). While both have similar ‘ideal’ weed characteristics, the difference in their success as invaders appears to lie in their relative ability to compete with other species in their target communities. C. canadensis allocates more resources to the production of flowers and seeds than C. sumatrensis, but has a lower competitive ability. Hence C. canadensis is more successful invading newly disturbed sites where seedlings do not compete for scarce resources. Some taxa are able to out-compete elements of the native flora which could lead to a homogenization of vegetation communities. An example would be the hottentot fig (Carpobrotus edulis), which originates from South Africa. It is a succulent creeper that can extend over large areas of dry coastal rocks, as in the western Mediterranean, where it out-competes the scattered native species (Mabberley and Placito 1993). A potential danger of invasive species is that, if successful in out-competing native flora, there will be an overall impoverishment and homogenization of vegetation communities and ecosystems, which might reduce their productivity and stability. However, it should not be assumed that all introduced species are potentially invasive. As demonstrated by the example of Conyza sumatrensis and C. canadensis, subtle biological differences are crucial in determining the successful establishment and eventual distribution of invasive taxa, which renders generalizations difficult to make. However, the degree of invasiveness may alter with global change.

The prickly-pear cactus (Opuntia ficus-indica) is a New World introduction. In the Iberian Peninsula it is invading old fields and there is a possibility that its expansion may be favoured by the predicted further increases in atmospheric carbon dioxide (Nobel and Decortazar 1990).

The Origin of Sclerophyllous Ecosystems The origin of sclerophyllous shrublands in the Mediterranean region is a contested issue (Allen 2003), with the debate polarizing between their spread through the region as a response to Holocene climatic changes and the actions of people. Reviewing the evidence for Holocene climate changes, Roberts et al. (2001a ) note that for the circumMediterranean region, there is a complex rather than simple pattern of change, potentially due to longitudinal shifts in atmospheric circulation. In other words, climate was not uniform across the entire area throughout any particular period, and nor were climate changes synchronous (see also Chapters 3, 4, and 9). Nevertheless they recognize marked climatic differences between the two halves of the Holocene. Pollen records from a number of sites across the region indicate the presence of deciduous, often oak-dominated, taxa prior to 6,000 (uncalibrated) years BP, suggesting wetter conditions than in the second half of the Holocene: for example at sites in the Iberian Peninsula (e.g. Sierra de Cebollera (Garcia et al. 2002)), in the northern Apennines of Italy (Lowe and Watson 1993), in Sicily (Sadori and Narcisi 2001), and in the eastern Mediterranean, such as central Anatolia in Turkey (Roberts et al. 2001b). However, climatically driven interpretations of the pollen record are equivocal, due in part to the topographical diversity of pollen-bearing sites across the region and their suitability for the preservation of an unambiguous climatic signal (Chapter 9). By 6,000 (uncalibrated) years BP reconstructions of European biomes point to the existence of a more humid and cooler mediterranean-type climate, which accords with general circulation model simulations predicting lower winter temperatures at the time, based on lower early Holocene winter insolation values (Prentice et al. 1996). From mid-Holocene times forest cover generally declined and pollen from deciduous oak was replaced by that of evergreen taxa, though comparisons of circum-Mediterranean pollen records suggest that the establishment of sclerophyllous communities was not necessarily synchronous, and that

Vegetation and Ecosystem Dynamics

emplacement was transitional from both east to west and south to north (Figure 7.13 and Chapter 9). This has been interpreted as a response to increasingly arid conditions from the mid-Holocene onwards and establishment of the modern mediterranean-type climate. In contrast to the argument for the climatically driven establishment of sclerophyllous shrublands, there are many advocates for an anthropogenic origin following the spread of agriculture from the Near East about 10,000 years ago (see, for example, Quézel 1999; Tomaselli 1981b). An anthropogenic explanation is consistent with the representation of sclerophyllous shrublands as degraded ecosystems as advocated, for example, by Turrill (1929) for the Balkans. He regarded forest as the climatic climax of most of the Mediterranean region, with maquis communities originating through over-exploitation of forest with further degradation to garrigue and steppe communities (Figure 7.8). Maquis as a degradation stage in succession, either progressive or retrogressive, was an influential idea in studies of Mediterranean ecosystems (for example, Tomaselli 1981b) and evergreen shrubland communities created by such disturbances have been regarded as disclimax or paraclimax communities which could revert to the climatic climax communities if disturbances ceased. The applicability of the concept of vegetation succession to mediterranean ecosystems is now being reassessed. In part this is due to the wider recognition that single pathway, unidirectional, and progressive succession is too simplistic (see e.g. Walker and del Moral 2003). For regions with a mediterraneantype climate competition for light may not be the primary driver of supposed successional changes; moisture stress is more likely to be a limiting factor. A tree might not have a competitive advantage over lower-growing shrubs (Blumler 1993). Thus woodland cover might not be the community most suited to the varied semi-arid climates of the Mediterranean. Detailed studies of modern woodlands show their structure and composition to be heavily influenced by human activities. Closed-canopy holm oak (Quercus ilex) forests, such as those found in Spain and southern France, have been managed since at least Neolithic times, generally as coppice plots (Terradas 1999). Charcoal production began in the Iron Age and was widespread until the nineteenth century. By the twentieth century it had mostly disappeared, except in a few localities in Portugal, North Africa, and the eastern Mediterranean (Blondel and Aronson 1999). Continued, and sustainable, exploitation of holm oak forests is believed to have promoted their survival and led to homogeniza-

223

tion of the woodlands at the expense of less-resistant taxa. Today, most forests bear evidence of cultivation and coppicing (Terradas 1999). Without intervention, closed-canopy holm oak forests are believed to be incapable of self-regeneration: few oak saplings grow in the understorey of established forest, despite the presence of seedlings. It seems likely that recruitment of seedlings to the sapling stage fails because they are not disturbancetolerant. Fire eliminates most seedlings and acorns, but does not kill the established trees as these are able to resprout (Retana et al. 1999). Production of acorns and their germination is then delayed until the resprouted individuals reach reproductive age. Resprouting individuals also out-compete seedlings that survive the fire, because of their already-existing root systems. A consequence of the millennial scale management of holm oak forests and occurrence of fire, is that new genotypes are not introduced into forest communities, causing gradual senescence and eventual degradation. However, in southern France there is, as yet, no sign of decline, despite continued existence since the Middle Ages.

Mediterranean Ecosystems in The Future Response to Elevated Levels of Atmospheric CO2 Contemporary climate changes are likely to have an effect on ecosystem dynamics in the Mediterranean. Increased levels of atmospheric carbon dioxide mean that a moderate increase in net primary productivity might be expected assuming plants are able to make more efficient use of available water (Mooney et al. 2001): experiments, under elevated concentrations of carbon dioxide, indicate that many species show increased efficiency of water use (Woodward et al. 1991). Taxa most likely to benefit are those which are late-growing species, shrubs, and leguminous plants that can make most use of increased availability of carbon dioxide, in other words, drought-tolerant nitrogenfixing species. Differential species response makes it difficult to assess the effects of elevated carbon dioxide. Not all plants will be able to take full advantage of increased carbon availability; early results from Californian mediterranean ecosystems tentatively suggested that in periodically stressful, low resource environments, the increase in net primary productivity might not be that great (Harley 1995). Experiments on monoliths of Mediterranean grasslands exposed to elevated carbon dioxide also

224

Harriet Allen

2

2

4.

5

0

7

7

9

7

2

7

10

10

4. 5

0

10

10

7

0

10

Isochron in thousands of years (Cal.)

?

Modern distribution Pre 10,000 years (Cal.)

0

500 km

Fig. 7.13. Spread of evergreen sclerophyllous vegetation across the Mediterranean region during the Holocene (after Roberts 1998) (see Chapters 4 and 9).

showed only a small increase in net primary productivity, probably as a result of low nutrient availability in the soils (Navas et al. 1995). In addition, while the initial effects of increased carbon dioxide are at the leaf level, with increased photosynthesis and decreased stomatal conductance, carbon allocation to other parts of a plant is poorly understood. There are also problems with scaling results from controlled environments, such as growth chambers, to ecosystems. Thus the precise effects of increased carbon dioxide on the functioning of Mediterranean ecosystems remain largely speculative.

Ecosystem Dynamics and Land Use Changes Divergent trends in land use changes are apparent in North Africa and the Near East and in the European Mediterranean (Barbero et al. 1990). In North Africa, population pressures are leading to major declines in forest cover, predominantly associated with collection of firewood, grazing, frequent use of fire, and expansion of cropland. Between 1965 and 1976 forest and shrubland decreased in area by 3 per cent while agricultural land increased by 5 per cent (Le Houérou 1981). By con-

trast, in southern Europe since the 1950s, rural depopulation and reorganization of agriculture following the availability of subsidies from the European Union, has resulted in recovery of forest land. In western Crete, for example, aerial photographs show an increase in forest cover of 75 per cent between 1945 and 1989 (Papanastasis and Kazaklis 1998), while in Mediterranean France, forest area increased by some 22 per cent between 1965 and 1976 (Le Houérou 1981). Among the areas most affected by land use changes are the uplands, at the economic margins of cultivation. Mention has already been made of the abandonment of terraces and the spread of maquis-type vegetation. Where the maquis taxa include especially flammable species, this could increase the risk of fire. It is feared that land use changes, such as those in the northern Mediterranean, will result in the homogenization of landscapes and an associated reduction in biodiversity, both in terms of species richness and functional attributes. Scenarios modelled for 2100 across global biomes suggest that Mediterranean biomes will experience a large loss of biodiversity because of their sensitivity to a range of factors affecting global change, especially land use changes (Sala et al. 2000). As argued

Vegetation and Ecosystem Dynamics

above, this has implications for ecosystem stability. Vegetation and ecosystem dynamics, together with soils, are the determinants of landscape heterogeneity and are therefore a vital component of habitats for natural history and recreation. For their long-term existence, Mediterranean ecosystems require continued disturbance, in the form of fires, grazing, and semi-traditional agricultural practices. Conservation measures should aim to maintain these, perhaps through extension of EU compensatory payments to farmers who follow environmentally beneficial practices and through the formal recognition of biodiversity-rich landscapes, such as the dehesas and montados of the Iberian Peninsula.

Acknowledgements The author would like to thank Roland Randall (Girton College, Cambridge), William Fletcher (University of Bordeaux) and Ian Lawson (University of Leeds) for their constructive comments, Ian Agnew (Deptertment of Geography, University of Cambridge) for the artwork, and Jamie Woodward for his encouragement.

References Allen, H. D. (2001), Mediterranean Ecogeography. Pearson, Harlow. (2003), Response of past and present Mediterranean ecosystems to environmental change, Progress in Physical Geography 27: 359–77. Randall, R. E., Amable, G. S., and Devereux, B. J. (2006), The impact of changing olive cultivation practices on the ground flora of olive groves in the Messara and Psiloritis regions, Crete, Greece. Land Degradation and Development 17: 249–73. Barbero, M., Bonin, G., Loisel, R., and Quézel, P. (1990), Changes and disturbances of forest ecosystems caused by human activities in the western part of the Mediterranean basin, Vegetatio 87: 151–73. Loisel, R., Quézel, P., Richardson, D. M., and Romane, F. (1998), Pines of the Mediterranean basin, in D. M. Richardson (ed.), Ecology and Biogeography of Pinus. Cambridge University Press, Cambridge, 153–70. Bergmeier, E. (1998), Are Cretan endemics threatened by grazing?, in V. P. Papanastasis (ed.), International Workshop on Ecological Basis of Livestock Grazing in Mediterranean Ecosystems, Thessaloniki, 90–3. Biger, G. and Liphschitz, N. (1991), The recent distribution of Pinus brutia: a reassessment based on dendroarchaeological and dendrohistorical evidence from Israel, The Holocene, 1: 157–61. Blondel, J. and Aronson, J. (1995), Biodiversity and ecosystem function in the Mediterranean basin: human and non-human determinants, in G. W. Davis and D. M. Richardson (eds.), Mediterranean-Type Ecosystems: The Function of Biodiversity. Springer, Berlin, 43–119. and Aronson, J. (1999), Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford.

225

Blumler, M. (1993), Successional patterns and landscape sensitivity in the Mediterranean and Near East, in D. S. G. Thomas and R. J. Alison (eds.), Landscape Sensitivity. John Wiley, Chichester, 287–305. Bottema, S. and Sarpaki, A. (2003), Environmental change in Crete: a 9000 year record of Holocene vegetation history and the effect of the Santorini eruption, The Holocene 13: 733–49. Christensen, N. L. (1985), Shrubland fire regimes and their evolutionary significance, in S. T. A. Pickett and P. S. White (eds.), The Ecology of Natural Disturbance and Patch Dynamics. Academic Press, Orlando, 85–100. (1994), The effects of fire on physical and chemical properties of soils in mediterranean-climate shrublands, in J. M. Moreno and W. C. Oechel (eds.), The Role of Fire in Mediterranean-Type Ecosystems. Springer, New York, 79–95. Clark, S. C., Puigdefábregas, J., and Woodward, I. (1998), Aspects of the ecology of the shrub-winter annual communities, in P. Mairota, J. B. Thornes, and N. Geeson (eds.), Atlas of Mediterranean Environments in Europe: The Desertification Context. John Wiley, Chichester, 44–7. Comps, B., Gömöry, D., Letouzey, J., Thiébaut, B., and Petit, R. J. (2001), Diverging trends between heterozogosity in the European beech, Genetics 157: 389–97. Cowling, R. M. (ed.) (1992), Fynbos: Nutrients, Fire and Diversity. Oxford University Press, Cape Town. Daget, P. (1980), Atlas d’Aréologie Périméditerranée. Institut de Botanique, Montpelier. di Castri, F. (1981), Mediterranean Shrublands of the World in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 1–52. Dimitrakopoulos, A. P. (2001), A statistical classification of mediterranean species based on their flammability components, International Journal of Wildland Fire 10: 113–18. and Papaioannou, K. K. (2001), Flammability assessment of Mediterranean forest fuels, Fire Technology 37: 143–52. Douglas, T. D., Kirkby, S. J., Critchley, R. W., and Park, G. J. (1994), Agricultural terrace abandonment in the Alpujarra, Andalucia, Spain, Land Degradation and Rehabilitation 5: 281–91. Garcia, M. J. G., Valiño, M. D., Rodríguez, A. M., and Zapata, M. B. R. (2002), Late-glacial and Holocene palaeoclimatic record from Sierra de Cebollera (northern Iberian Range, Spain), Quaternary International 93/4: 13–18. Grove, A. T. and Rackham, O. (2001), The Nature of Mediterranean Europe An Ecological History. Yale University Press, New Haven. Groves, R. H. and di Castri, F. (eds.) (1991), Biogeography of Mediterranean Invasions. Cambridge University Press, Cambridge. Harley, P. C. (1995), Modeling leaf level effects of elevated CO2 on Mediterranean sclerophylls, in J. M. Moreno and W. C. Oechel (eds.), Global Change and Mediterranean-Type Ecosystems. Springer, New York, 35–57. Harrison, R. J. (1996), Arboriculture in southwest Europe: dehesas as managed woodlands, in D. Harris (ed.), The Origins and Spread of Agriculture and Pastoralism in Eurasia. UCL, London, 363–7. Hector, A., Schmid, B., and 32 others (1999), Plant diversity and productivity experiments in European grasslands, Science 286: 1123–7. Hewitt, G. (2000), The genetic legacy of the Quaternary ice ages, Nature 405: 907–13.

226

Harriet Allen

Joffre, R. (1992), The Dehesa: does this complex ecological system have a future?, in A. Teller, P. Mathy, and J. N. R. Jeffries (eds.), Responses of Forest Ecosystems to Environmental Changes (London), 381–8. Klaus, W. (1989), Mediterranean Pines and their History, Plant Systematics and Evolution 162: 133–63. Legge, T. (1996), The Beginning of Caprine Domestication in Southwest Asia, in D. Harris (ed.), The Origins and Spread of Agriculture and Pastoralism in Eurasia. Elsevier, London, 238–62. Le Houérou, H. N. (1981), Impact of man and his animals on Mediterranean vegetation, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-type Shrublands. Elsevier, Amsterdam, 479–521. (1990), Global change: vegetation, ecosystems and land use in the southern Mediterranean basin in the mid twenty-first century, Israel Journal of Botany 39: 481–508. Lowe, J. J. and Watson, C. (1993), Lateglacial and Early Holocene pollen stratigraphy of the Northern Apennines, Italy, Quaternary Science Reviews 12: 727–38. Loumou, A. and Giourga, C. (2003), Olive groves: the life and identity of the Mediterranean, Agriculture & Human Values 20: 87–95. Mabberley, D. J., and Placito, P. J. (1993), Algarve Plants and Landscape. Oxford University Press, Oxford. McNeill, J. R. (1992), The Mountains of the Mediterranean World. Cambridge University Press, Cambridge. Margaris, N. S. (1977), Physiological and biochemical observations in seasonal dimorphic leaves of Sarcopoterium spinosum and Phlomis fruticosa, Oecologia Plantarum 12: 343–50. (1981), Adaptive Strategies in Plants dominating Mediterranean-type Ecosystems, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 309–15. Koutsidou, E., and Giouga, C. (1996), Changes in traditional Mediterranean land use systems, in J. C. Brandt and J. B. Thornes (eds.), Mediterranean Desertification and Land Use. John Wiley, Chichester, 29–42. Médail, F. and Quézel, P. (1997), Hot-spots analysis for conservation of biodiversity in the Mediterranean basin, Annals of the Missouri Botanical Garden 84: 112–27. Mooney, H. A. and Dunn, E. L. (1970), Convergent evolution of Mediterranean-climate evergreen sclerophyllous shrubs, Evolution 24: 292–303. Kalin Arroyo, M. T., Bond, A. J., Candell, J., Hobbs, R. J., Lavorel, S., and Neilson, R. P. (2001), Mediterranean-climate ecosystems, in F. S. Chapin III, O. E. Sala, and E. HuberSannwald (eds.), Global Diversity in a Changing Environment:. Scenarios for the Twentyfirst Century. Springer, New York, 157–99. Navas, M.-L., Guillerm, J.-L., Fabreguettes, J., and Roy, J. (1995), The influence of elevated CO2 on community structure, biomass and carbon balance of Mediterranean old-field microcosms, Global Change Biology 1: 325–35. Naveh, Z. (1974), The Effects of fire in the Mediterranean region, in T. T. Kozlowski and C. E. Ahlgren (eds.), Fire and Ecosystems. Academic Press, New York, 401–37. (1975), The evolutionary significance of fire in the Mediterranean region, Vegetatio 9: 199–206. (1994), The role of fire and its management in the conservation of Mediterranean ecosystems and landscapes, in J. M. Moreno and W. C. Oechel (eds.), The Role of Fire in Mediterranean-Type Ecosystems. Springer, New York, 163–85.

(1995), Conservation, restoration and research priorities for Mediterranean uplands threatened by climate change, in J. M. Moreno and W. C. Oechel (eds.), Global Change and Mediterranean-Type Ecosystems. Springer, New York, 482–507. and Whittaker, R. H. (1979), Structural and floristic diversity of shrublands and woodlands in northern Israel and other Mediterranean areas, Vegetatio, 41: 171–90. Nobel, P. S. and Decortazar, V. G. (1991), Growth and predicted productivity of Opuntia ficus-indica for current and elevated carbon dioxide, Agronomy Journal 83: 224–30. Ojeda, F., Arroyo, J., and Marañon, T. (1995), Biodiversity components and conservation of Mediterranean heathlands in Southern Spain, Biological Conservation 72: 61–72. Papanastasis, V. P. and Kazaklis, A. (1998), Land-Use changes and conflicts in the Mediterranean-type ecosystems of Western Crete, in P. W. Rundel, G. Montenegro, and F. M. Jaksic (eds.), Landscape Disturbance and Biodiversity in MediterraneanType Ecosystems. Springer, Berlin, 141–54. Pausas, J. G. and Vallejo, V. R. (1999), The role of fire in European Mediterranean ecosystems, in E. Chuvieco (ed.), Remote Sensing of Large Wildfires in the European Mediterranean Basin. Springer, Berlin, 3–16. Perevolotsky, A. and Haimov, Y (1992), The effect of thinning and goat browsing of the structure and development of Mediterranean woodland in Israel, Forest Ecology and Management, 49: 61–74. Phitos, D., Strid, A., Snogerup, S., and Greuter, W. (eds.) (1995), The Red Data Book of Rare and Threatened Plants in Greece. WWF, Athens. Polunin, O. and Smythies, B. E. (1973), Flowers of South-West Europe. Oxford University Press, Oxford. Prentice, I. C., Guiot, J., Huntley, B., Jolly, D., and Cheddadi, R. (1996), Reconstructing biomes from Palaeoecological data: a general method and its application to european pollen data at 0 and 6 ka, Climate Dynamics 12: 185–94. Quézel, P. (1981a ), The study of plant groupings in the countries surrounding the Mediterranean: some methodological aspects, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 87–93. (1981b), Floristic composition and phytosociological structure of sclerophyllous matorral around the Mediterranean, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 107–21. (1985), Definition of the Mediterranean region and the origin of its flora, in C. Gómez-Campo (ed.), Plant Conservation in the Mediterranean Area. Kluwer, Dordrecht, 8–24. (1999), Les grandes structures de végétation en région méditerranéenne: facteurs déterminants dans leur mise en place post-glaciare, Geobios 32: 19–32. Rackham, O. and Moody. J. (1996), The Making of the Cretan Landscape. Manchester University Press, Manchester. Raven, P. H. (1973), The evolution of Mediterranean floras, in F. di Castri and H. A. Mooney (eds.), Mediterranean Type Ecosystems: Origin and Structure. Springer, Berlin, 213–24. Retana J., Espelta, J. M., Gracia, M., and Riba, M. (1999), Seedling recruitment, in F. Rodà, J. Retana, C. A. Gracia, and J. Bellot (eds.), Ecology of Mediterranean Evergreen Oak Forests. Springer, Berlin, 89–103. Roberts, C. N. (1998), The Holocene: An Environmental History. Blackwell, Oxford.

Vegetation and Ecosystem Dynamics Roberts, N., Meadows, M. E., and Dodson, J. R. (2001a ), The history of Mediterranean-type environments: climate, culture and landscape, The Holocene 11: 631–4. Reed, J. M., Leng, M. J., and 8 others. (2001b), The tempo of Holocene climatic change in the Eastern Mediterranean region: new high-resolution crater-lake sediment data from Central Turkey, The Holocene 11: 721–36. Ross, J. D. and Sombrero, C. (1991), Environmental control of essential oil production in Mediterranean plants, in J. B. Haborne and F. A. Tomas-Barberan (eds.), Ecological Chemistry and Biochemistry of Plant Terpinoids. Oxford University Press, Oxford, 83–94. Ruíz, M. and Ruíz, J. P. (1986), Ecological history of transhumance in Spain, Biological Conservation 37: 73–86. Sadori, L. and Narcisi, B. (2001), The postglacial record of environmental history from Lago di Pergusa, Sicily, The Holocene 11: 655–72. Sala, O. E., Chapin III, F. S., and 17 others (2000), Global biodiversity scenarios for the year 2100, Science 287: 1770–4. Schulze, E.-D. and Mooney, H. A. (1994), Ecosystem function of biodiversity: a summary, in E.-D. Schulze and H. A. Mooney (eds.), Biodiversity and Ecosystem Function. Springer, Berlin. Sfikas, G. (2002), Wildflowers of Crete. Efstathiadis, Athens. Specht, R. (1979), Ecosystems of the World, ix. Heathland and Related Shrublands. Elsevier, Amsterdam. Spehn, E. M., Scherer-Lorenzen, M., Schmid, B., Hector, A., Caldeira, M. C., Dimitrakopoulos, P. G., Finn, J. A., Jumpponen, A., O’Donovan, G., Pereira, J. S., Schulze, E.-D., Troumbis, A. Y., and Körner, C. (2000), The role of legumes as a component of biodiversity in a cross-European study of grassland biomass nitrogen, Oikos 98: 205–18. Suc, J.-P. (1984), Origin and evolution of the Mediterranean vegetation and climate in Europe, Nature 307: 429–32. Terradas, J. (1999), Holm oak and holm oak forests: an introduction, in F. Rodà, J. Retana, C. A. Gracia, and Bellot, J. (eds.), Ecology of Mediterranean Evergreen Oak Forests. Springer, Berlin, 3–14. Thébaud, C., Finzi, A., Affre, L., Debussche, M., and Escarre, J. (1996), Assessing why two introduced conyza differ in their ability to invade Mediterranean old fields, Ecology 77: 791– 804. Thompson, J. D. (1999), Population differentiation in Mediterranean plants: insights into colonization history and evolution and conservation of endemic species, Heredity 82: 229–36. Thornes, J. B. (1998), Results and prospect, in P. Mairota, J. B. Thornes, and N. Geeson (eds.), Atlas of Mediterranean Environments in Europe: The Desertification Context. John Wiley, Chichester, 162–3. Tilman, D. (2000), Overview: Causes, Consequences and Ethics of Biodiversity, Nature 405: 208–11. Tomaselli, R. (1981a ), Main physiognomic types and geographic distribution of shrub systems related to Mediterranean

227

climates, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 95–106. Tomaselli, R. (1981b), Relations with other ecosystems: temperate evergreen forests, Mediterranean coniferous forests, savannahs, steppes and desert shrublands, in F. di Castri, D. W. Goodall, and R. Specht (eds.), Ecosystems of the World, xi. Mediterranean-Type Shrublands. Elsevier, Amsterdam, 123–30. Trabaud, L. (1987), Natural and prescribed fire: survival strategies of plants and equilibrium in Mediterranean ecosystems, in J. D. Tenhunen, F. M. Caterino, and O. L. Lange (eds.), Plant Response to Stress. Springer, Berlin, 607–21. (1994), Postfire plant community dynamics in the Mediterranean basin, in J. M. Moreno and W. C. Oechel (eds.), The Role of Fire in Mediterranean-Type Ecosystems. Springer, New York, 1–15. Christensen, N. L., and Gill, A. M. (1993), Historical biogeography of fire in temperate and mediterranean ecosystems, in P. J. Crutzen and J. G. Goldammer (eds.), Fire in the Environment: Its Ecological and Atmospheric Importance. John Wiley, New York, 277–95. Troumbis, A. Y., Dimitrakopoulos, P. G., Siamantziouras, A.S. D., and Memtsas, D. (2000), Hidden diversity and productivity patterns in mixed Mediterranean grasslands, Oikos 90: 549–59. Turrill, W. B. (1929), The Plant Life of the Balkan Peninsula. Oxford University Press, Oxford. Tzedakis, P. G. (1993), Long-term tree populations in Northwest Greece through multiple Quaternary climatic cycles, Nature, 364: 437–40. Uerpmann, H. P. (1996), Animal domestication—accident or intention?, in D. Harris (ed.), The Origins and Spread of Agriculture and Pastoralism in Eurasia. UCL, London, 238–62. Vogiatzakis, I. N., Griffiths, G. H., and Mannion, A. M. (2003), Environmental factors and vegetation composition, Lefka Ori Massif, Crete, S. Aegean, Global Ecology and Biogeography 12: 131–46. Walker, L. R. and del Moral, R. (2003). Primary Succession and Ecosystem Rehabilitation. Cambridge University Press, Cambridge. Willis, K. J. and Whittaker, R. J. (2002), Species diversity-scale matters, Science 295: 1245–347. Woodward, F. I., Thompson, G. B., and McKee, I. F. (1991), The effects of elevated concentrations of carbon dioxide on individual plants, populations, communities and ecosystems, Annals of Botany 67: 23–38. Zamora, J., Verdúa, J. R., and Galantea, E. (2007), Species richness in Mediterranean agroecosystems: spatial and temporal analysis for biodiversity conservation. Biological Conservation 134: 113–21. Zohary, M. (1962), Plant Life in Palestine. Ronald, New York.

This chapter should be cited as follows Allen, H D. (2009), Vegetation and ecosystem dynamics, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 203–227.

This page intentionally left blank

8

Hydrology, River Regimes, and Sediment Yield John Thornes, Francisco López-Bermúdez, and Jamie Woodward

Introduction In comparison to the rest of Europe, Africa, and Asia, most rivers arising and flowing within the Mediterranean watershed typically drain small catchments with mountainous headwaters (Figure 8.1). The hydrology of Mediterranean catchments is strongly influenced by the seasonal distribution of precipitation, catchment geology, vegetation type and extent, and the geomorphology of the slope and channel systems. It is important to appreciate, as the preceding chapters have shown, that the area draining to the Mediterranean Sea is large and enormously variable in terms of the key controls on catchment hydrology outlined above, and it is therefore not possible to define, in hydrological terms, a strict single Mediterranean river type. However, river regimes across the basin do have a marked seasonality that is largely controlled by the climate system (Chapter 3) and, in most basins, the dominant flows occur in winter—but autumn and spring runoff is also important in many areas. These patterns reflect the general water balance of the basin as a whole, but there are key geographical patterns in catchment hydrology and sediment yield and a marked contrast is evident between the more humid north and the semi-arid south and east (Struglia et al. 2004; Chapter 21). Also, because of the long history of vegetation and hillslope modification by human activity and the more recent and widespread implementation of water resource management projects, there are almost no natural river regimes in the Mediterranean region, especially in the middle and lower reaches of river catchments (Cudennec et al. 2007).

Runoff generation on hillslopes in the Mediterranean is very closely related to rainfall intensities and land surface properties as discussed in Chapter 6. While this is probably true of most catchments, runoff generation in the Mediterranean is very sensitive to vegetation cover because of the seasonal dynamics of rainfall and the role played by extreme events. The cumulative effect of these characteristics is a specific set of management problems and restoration issues and, although these are rather different in the various socio-political regimes of the region, it can be argued that they are in many ways unique to Mediterranean catchments.

The Water Balance and Mediterranean River Regimes The Annual Water Budget The broad water balance for the Mediterranean basin as a whole is summarized in Figure 8.2 which is based on the work of Jean Margat (Chapter 21) and published in Grenon and Batisse (1989). This shows mean values for annual water fluxes in the region in km3 (1 cubic kilometre = 1 billion m3 ). It is important to realize that these values will vary from year to year and they mask very significant spatial variations in catchment hydrology as we explain below. Nonetheless Figure 8.2 provides a very useful illustration of the magnitude of the annual water exchanges in the region between the atmosphere, the terrestrial catchment systems, and the Mediterranean Sea itself. Annual precipitation input to the Mediterranean ‘catchment’ is of the order of

230

John Thornes et al.

Fig. 8.1. An upland river catchment in the mountains of north-west Greece (photo: Jamie Woodward).

1,070 km3 each year with total evaporation losses of around 685 km3 per year. Taking into account contributions from larger river systems (such as the Nile and the Rhône that drain areas beyond the Mediterranean region), groundwater flows, and human water uses, this

budget gives an annual terrestrial water yield to the Mediterranean Sea of about 477 km3 . This component of the water balance in relation to the annual exchange of water between the Mediterranean Sea and the Atlantic Ocean is discussed in Chapter 2 and all components of

Hydrology, River Regimes, and Sediment Yield

231

Fig. 8.2. The water balance of the Mediterranean region showing the major fluxes (in km3 per year) between the main components of the hydrological cycle (modified from Grenon and Batisse 1989, and based on the work of Jean Margat). Values in brackets represent earlier estimates of these components.

232

John Thornes et al.

19.5

60

6 58.5 6 159

8.5

3

21

1.1

1.4

5 53

3 2

63.5

0.5 6

0

500 km

Fig. 8.3. Total annual runoff from river basins (shaded area) in each country bordering the Mediterranean Sea. The runoff values are given in km3 per year. The runoff volumes from river catchments in Libya, Israel, and Palestine (shown as Â) are negligible at this scale. The value given here for the Nile reflects the impact of the High Aswan Dam (modified from Grenon and Batisse 1989).

this water budget are expected to change significantly in coming decades given the predictions for climate change in the region presented in Chapter 3. An examination of annual catchment runoff to the Mediterranean Sea on a country by country basis shows large spatial variations in river discharge. Figure 8.3 shows the average annual flows (km3 ) as they were in 1985 and reported by Grenon and Batisse (1989) as part of the Blue Plan initiative. The main feature here is the marked contrast between the perennially flowing rivers on the north side of the basin and the very small contributions made by the strongly seasonal and ephemeral river catchments on the south side of the basin and in the Levant. In fact, the total contribution from river systems in Morocco, Algeria, and Tunisia (12.6 km3 per year) is almost matched by the combined runoff from just Corsica and Sardinia. Fluvial runoff to the Mediterranean Sea is dominated by river systems in Italy, Turkey, France, Greece, and the countries of the former Yugoslavia. Flows from the Nile have fallen dramatically since the construction of the High Aswan Dam and major irrigation schemes (Woodward et al. 2007).

The total annual flow of Mediterranean rivers depends on the annual water balance partition in their catchments. This has been studied and modelled by Eagleson (1978) and is depicted in Figure 8.4. This shows how annual rainfall input is divided into evapotranspiration loss from soil moisture, seepage loss to groundwater, and rainfall excess (that becomes river runoff). As mean annual precipitation increases from left to right along the horizontal axis, the rainfall excess increases, as shown in the shaded area. Thus the vertical line (1) represents the water-balance of rivers from the semi-arid and arid parts of the Mediterranean, such as those of south-east Spain, North Africa, and Israel. Vertical line (2) is the lower boundary of the sub-humid water balance regimes, found on the northern edges of the Mediterranean basin where rivers have sufficient water to flow the year round, and the soil moisture is sufficient to support a ‘good’ ground cover of vegetation. Finally, line (3) represents the ‘humid’ Mediterranean regimes of rivers with wooded mountain catchments (Chapters 3 and 7) and, of course, there are distinctive hydrological balances for special cases such as wetland systems (Chapter 9) and karst landscapes (Chapter 10).

Hydrology, River Regimes, and Sediment Yield

233

Fig. 8.4. Eagleson’s decomposition of the annual water balance for catchments in different climatic settings. The shaded area is the ‘rainfall excess’ or the runoff to rivers. See text for further explanation (modified from Eagleson 1978). The groundwater runoff component is especially important in many Mediterranean catchments because of the widespread occurrence and thickness of limestone rocks (Chapter 10).

Seasonal Distribution of River Flows Although all Mediterranean rivers have flows that have been greatly affected by human activity, the monthly pattern of flows or river regimes still reflects the broadly biophysical controls of climate and to a lesser extent groundwater storage. These climatic effects produce the major spatial distributions of vegetation types (Chapter 7) which combine to form key controls on land degradation as discussed in Chapter 20. The seasonal distribution of river flows is controlled mainly by the seasonal distribution of rainfall as broadly shown in Figure 8.5. This diagram also provides an approximation of the range of river regime types across the Mediterranean region. The great allogenic rivers such as the Rhône and the Nile have headwaters outside the Mediterranean winter rainfall regime and have hybrid regimes and the Rhône is considered in more detail below. Typically the Mediterranean has rains in winter and autumn (Chapter 3) with the middle and lower reaches of many rivers being dry in the summer (Figure 8.6). This is characteristic of much of the Maghreb, the Levant, southern Greece, southern Italy, the very south-east and south of the Iberian Peninsula and most of Portugal. Most of the Iberian Peninsula, Italy, and the western parts of the Balkan Peninsula also receive significant autumn rains. Only the continental parts of the Balkan states and Macedonia have rainfall

all year round. Monthly flow series for eight river gauging stations across the Mediterranean region (and one in the headwaters of the Rhône) are given in Figure 8.7. The histograms show the range of river regimes evident in the Mediterranean moving clockwise (a) to (i) around the basin (Figure 8.5) and include river systems in France, Italy, Albania, Turkey, Cyprus, Israel, Algeria, and Morocco. All the stations (b) to (i) show a marked summer decrease in river flows with the channel bed of the Vasilikos River (151 km2 ) in Cyprus drying out completely for five months from July to November. The flow data for the rivers in the east and south of the Mediterranean (e) to (i) highlight the importance of winter and spring precipitation in these catchments. The Rhône at Chancy, on the French–Swiss border, has a mean annual discharge of 342 m3 s−1 . The minimum and peak recorded flows at this station in the period January 1965 to December 1982 were 134 and 740 m3 s−1 respectively. This regime reflects the essentially Alpine character of the headwaters, with a strong spring and summer snowmelt. By the time the river reaches Beaucaire in the Garrigue, the Mediterranean regime is already evident. The catchment area here is almost 100,000 km2 , about nine times that at Chancy. The mean discharge is >1,600 m3 s−1 with minimum and maximum flows of 420 and 5,077 m3 s−1 respectively. The strong dip

234

John Thornes et al.

Fig. 8.5. Rainfall regimes in the Mediterranean region (modified from Huttary 1955). The solid line is the watershed of rivers that drain to the Mediterranean Sea. The letters (a to i) mark the locations of the gauging stations in Figure 8.7.

in the summer and early autumn months (July to October) is quite characteristic of the Mediterranean regime. Snowmelt makes an important contribution to many headwater river systems in the Mediterranean region—especially where the mountains rise above 2,000 m (Chapter 12). The regime for Wadi Moulouya in Morocco—at the Dar el Caid station—shows a more pronounced Mediterranean regime. Here the annual mean discharge is 20 m3 s−1 with minimum and maximum peak flows of 0 and 327 m3 s−1 respectively. August has the lowest average flow (3.6 m3 s−1 ) and April, the maximum, showing the spring rainfall effect (Figure 8.7). The catchment of the Wadi Moulouya at Dar el Caid (24,422 km2 ) is similar in size to the Menderes River at Soke (23,889 km2 ) in Turkey yet, due to the greater aridity of the local climate, its mean annual discharge is only about one fifth of the Menderes (Figure 8.7). For large parts of the semi-arid Mediterranean, channels may be without flow for most of the summer but commonly restarting with the onset of autumn storms. Surface channel flows in these rivers normally cease in early May, though flow will often continue beneath the gravel beds and may be used for irrigation throughout the summer months. Many channels are truly

ephemeral and convey flows only during major storm events (Cudennec et al. 2007)). In flash floods, water drains into the channel beds so the flows may only be evident in upland bedrock reaches and then disappear in the middle and downstream reaches where the channels are formed in thick coarse-grained gravely sediments (Thornes 1977; Butcher and Thornes 1978; Shannon et al. 2002). A key problem in many parts of the semiarid Mediterranean is the absence of reliable, long-term records of river flows.

Human Modification of Mediterranean Catchment Hydrology The human history of the Mediterranean lands has been strongly moulded by its rivers (Lewin et al. 1995; Grove and Rackham 2003; Chapter 11) and among all the natural ecosystems, rivers and floodplains are the most intensively used by Mediterranean people. The floodplains have formed important corridors for the migration of plants, animals, and people and have also provided some of the most fertile alluvial soils in the

Hydrology, River Regimes, and Sediment Yield

235

Fig. 8.6. The seasonally dry gravel-bed channel of the Voidomatis River upstream of the Vikos Gorge in north-west Greece (photo: Jamie Woodward). Note the high flow stage marks on the bedrock wall.

world, on which intensive irrigated agriculture has been established. Around them, complex legal systems for their use have developed. As in other parts of the world, the human modification of river flows and channels has taken place both directly and indirectly. Indirect changes are those which change runoff generation and sediment supply at the basin scale—mainly through changes in land use. Direct changes are those which alter the flow regimes, sediment stores, and planform of the channel, usually by engineering works related to water resource development and flood protection or alluvial gravel extraction (e.g. Christopoulos 1998; Nicholas et al. 1999; Kapsimalis et al. 2005) (Chapter 11). What sets the Mediterranean region apart from other areas of the world, however, is the long history and widespread occurrence of both direct and indirect modifications to catchment hydrology and river channel systems (VitaFinzi 1978; Macklin et al. 1995).

Indirect Changes to Catchment Systems Vegetation Cover and Runoff Dynamics Elwell and Stocking (1976) showed that runoff from experimental plots was reduced as the vegetation cover was increased and the curve is exponential (Figure 8.8). On the horizontal axis is the percentage vegetation cover in plan view. On the vertical axis is the amount of runoff as a percentage of bare soil. This parameter is the result of many complex factors, as discussed later in this chapter, most prominently rainfall intensity, basin morphology, and soil characteristics, such as infiltration capacity and crusting. These experiments explain the earlier discoveries of American workers and the speculations of the ancient Greeks (Grove and Rackham 2003) and have been reconfirmed in many studies since, in

(a) Rhône at Chancy, Switzerland Catchment = 10,299 km2 Mean discharge = 341.61 m3 s-1

(b) Rhône at Beaucaire, France Catchment = 95,590 km2 Mean discharge = 1,694.95 m3 s-1

(c) Tiber at Rome, Italy Catchment = 16,545 km2 Mean discharge = 231.13 m3 s-1

2000

300

Discharge (m3 s-1)

600 500 400

1500 200

300

1000

200 100 0

J FM AM J J A S O N D

(d) Drin at Kalimash, Albania Catchment = 12,368 km2 Mean discharge = 219.43 m3 s-1

Discharge (m3 s-1)

300

0

J FM AM J J A S O N D

(e) Menderes at Soke, Turkey Catchment = 23,889 km2 Mean discharge = 101.24 m3 s-1 250 200

0

J FM AM J J A S O N D

(f) Vasilikos at Zyyi, Cyprus Catchment = 151 km2 Mean discharge = 0.16 m3 s-1 0.8

0.6

150

200

0.4 100 100

0.2

50 0

0 J FM AM J J A S O N D

(g) Jordan at Obstacle Bridge, Israel Catchment = 1,376 km2 Mean discharge = 18.05 m3 s-1

Discharge (m3 s-1)

100

500

0 J FM AM J J A S O N D

J FM AM J J A S O N D (h) Cheliff at Sidi Belatar, Algeria Catchment = 43,750 km2 Mean discharge = 27.21 m3 s-1 100

(i) Moulouya at Dar el Caid, Morocco Catchment = 24,422 km2 Mean discharge = 21.15 m3 s-1 40

30 75

30 20

50

10

0

20

25

J FM AM J J A S O N D

0

10

J FM AM J J A S O N D

0

J FM AM J J A S O N D

Fig. 8.7. Average monthly flows for rivers around the Mediterranean basin based on data from various sources cited in the text. Catchment sizes vary greatly, but note that the water fluxes for rivers draining the northern part of the basin, (a) to (e), are at least an order of magnitude higher that those draining the east and south of the basin, (f) to (i).

Hydrology, River Regimes, and Sediment Yield

Deforestation and Catchment Hydrology

80

60

Runoff (l m-1)

237

40

20

0 0

20 40 60 Vegetation canopy cover (%)

80

Fig. 8.8. The relationship between relative runoff and vegetation canopy cover according to Elwell and Stocking (1976).

theoretical (Carson and Kirkby 1972), experimental (Francis and Thornes 1990), and field investigations (Bochet et al. 2006). In the last paper, the authors found that, for the three species of Mediterranean shrub considered, whereas total vegetation canopy and litter cover reduced soil loss and runoff rates, soil loss and runoff increased with plant volume. A very careful and detailed study of the effect of canopy structure on rainfall intensity for three common shrub species of the western Mediterranean (Anthyllis cytisoides, Stipa tenacissima, and Retama sphaerocarpa) was undertaken in laboratory simulations by GarciaOrtiz (2006). The species are shown in Figure 8.9. She found significant differences between species in canopy drip and through-fall volumes (as a percentage of the precipitation) and a correspondence between the main partitioning pathways and the root system type of the species was identified. Species with a deep root system were more efficient at redistributing rain by stem flow whereas species with a shallow root system redistribute rain mainly as canopy drip (Stipa tenacissima). Retama sphaerocarpa is the species with the least canopy storage, while Anthyllis cytisoides has the highest. The effects of rainfall intensity could not be fully explored and there is a need here for further research. It appears that Darwinian selection has led to plant morphologies that efficiently use the available rainfall to ensure their survival and may explain why these species are so abundant.

It follows that changes in the vegetation cover will transform the runoff regimes of Mediterranean catchments. This proposition is theoretically and empirically sound, but the results of studies of deforestation have not always been so clear. This is partly because of the wide variations in what is regarded as ‘forest’, because deforestation covers a wide variety of patterns of cutting within a catchment and because post-deforestation recovery depends on the sequence of weather in the years following cutting. These factors have been studied in detail by Obando (2002) in a semi-arid Mediterranean environment and the results suggest caution is needed in overenthusiastic endorsement of reafforestation as a strategy for reducing runoff. Chirino et al. (2006) concluded from their study of the effects on runoff of a thirty-year-old Aleppo pine plantation in south-east Spain that the conservation of natural communities or the restoration of semi-arid ecosystems with native shrub species can be as effective as reafforestation in managing runoff and erosion as suggested by Francis and Thornes (1994). The same care needs to be exercised when claiming major impacts on catchment runoff resulting from fire and grazing. A large number of studies have been carried out on the effects of fire on water and sediment yield in the Mediterranean (Wainwright and Thornes 2003). Most of these results demonstrate an increase of at least one order of magnitude in both runoff and sediment yield. Variations in the intensity of the burn yield different results, as shown by Inbar (1992). The impacts of fire are largely in changing the infiltration properties of soils by changing their hydrophobic properties. These changes have to be balanced against the positive benefits of both reafforestation and grazing. As with soil erosion, the issue is simply not clear cut, nor are the conclusions uniform (Chapter 20). Chapter 19 deals with the wildfire hazard more generally in the Mediterranean.

Cultivation and Catchment Hydrology There is less doubt about the impacts of cultivation on catchment hydrology, though still some scope for disagreement. Rapid runoff from hillslope and channel systems is a major catchment management issue across the Mediterranean region (Figure 8.10a). Wainwright (1996) examined the controls on runoff in the September 1992 flood in southern France in the basin of the River Ouvèze. Although the flooding was principally a result of the high rainfall intensities that occurred

Fig. 8.9. The species used by Garcia-Ortiz in her study of rainfall partitioning by different Mediterranean plants. (a) Anthyllis Cytisoides (b) Retama retamo (c) Stipa tenecissima. The Retama is about 2 m high (photos: John Thornes).

Hydrology, River Regimes, and Sediment Yield

239

that it is unimportant. It is that the activity is so diverse and its intensity so variable through time that it is difficult to predict the outcomes of these human modifications on catchment runoff and river regimes in Mediterranean environments. Nevertheless, because the effects of vegetation on runoff are reasonably well understood, they provide us with a tool for managing grazing in a way that optimizes runoff and sediment yield by assigning areas for grazing use (Thornes 2007).

Direct Changes to Catchment Systems Terraces and Water Harvesting

Fig. 8.10. Flood flows and erosion in Mediterranean catchments. (a) Runoff during a flash flood in Tunisia (photo: Ian Foster). (b) Deep rill erosion on sloping agricultural land in the western Peloponnese, Greece (photo: Jamie Woodward).

in the catchment, experimental and modelling evidence suggest that the severe flooding could be explained by runoff production in the agricultural areas on the lower slopes and terraces immediately upstream of the village of Vaison, as well as from the badland areas which are concentrated more in the upland areas of the catchment. He indicated that cultivation practices encouraged flow concentrations, accentuating the links between slope and channel. An experimental study, using digital simulation, by Kirkby et al. (2002) indicated how steepness of the terrain and the direction of ploughing, relative to the main slope direction, influenced the volume of runoff. Figure 8.10b shows how heavy rainfall and overland flow on bare and steeply sloping agricultural land in the Alfios River basin of the western Peloponnese can lead to the formation of deep (>60 cm) rills and the loss of valuable topsoil. As with deforestation and cultivation, the argument about the impacts of grazing on catchment runoff is not

One of the most ancient and widely practised direct water control measures in the Mediterranean has been the construction of agricultural terraces (Figure 8.11) (Grove and Rackham 2003; Chapters 6 and 7). This is a skilful procedure. The terraces have to be easy to cultivate with draft animals (oxen or mules), accessible, and fairly flat, and yet designed to ensure that the water is conducted away safely without undue erosion. Water infiltrates into the terraces and, if hydraulic pressure builds up inside the terrace, it can lead to piping which in turn can produce deep shafts big enough to swallow animals and machines. In North Africa and the Levant, the management of catchments has become a fine art, in order to conserve water for agriculture (Evenari et al. 1982). This procedure is called water harvesting. In carefully selected areas, vegetation is removed and the soil ‘crusted’ by trampling or hardening. Then the runoff from these crusted surfaces is conveyed by lined ditches into tanks, cisterns (Figure 8.12), or aljibe (tanks with rounded roofs that collect evaporated water) to be used for crop irrigation. This practice was well established in Roman times and used by Arabic farmers in the Maghrebian states and Iberia (Gómez-Espin 2004). The modern equivalent of the ancient stone-fronted terraces is extensive contour benching over large areas, usually by bulldozers. These are direct conservation measures for runoff on the hillslopes and may be used for flood protection.

Check Dams and Reservoir Development Direct measures in the channels range from small check dams, a few metres high, to major reservoirs (Figure 8.13a, b). Check dams serve multiple purposes. In addition to reducing the amount and velocity of runoff from headwater channels, they also store transported

240

John Thornes et al.

Fig. 8.11. Stone-walled terraces on hillslopes near Campanet in central Majorca (photo: Jamie Woodward).

sediment and provide flat areas for cultivation, into which water quickly and easily infiltrates. Another small, direct control on runoff in channels that dates back to the ancient water-control era of the western Mediterranean is the boquera system. Walls are constructed in the channel floor to lead flood water into fields, or boqueras, on either side of the channel where water is stored, as soil moisture, to provide for cereal and tree crop cultivation. Mill sites also abound in channel systems and the power provided can then lift irrigation water into canal systems that may redistribute it across hundreds of hectares. Other minor direct features include tunnels constructed beneath river beds—the water infiltrates into the gravels and then into permeable brick-lined tunnels that may convey it as much as 30 km to provide town water supplies. Many inter-basin water transfers were introduced in the Mediterranean region during the Roman Period as the expansion of urban settlements increased the demand for water supplies. In southern Epirus (north-west Greece), for example, the city of Nikopolis, founded by Octavius Augustus in 30 BC, required the construction of a major aqueduct (50 km in length) that brought large volumes of water from the Louros Valley to the north to the 300,000

residents of the city (Mavromati and Chryssaidis 2007; Vita-Finzi 1978). Above all, the runoff of rivers has been modified by the great reservoirs the construction of which commenced at the end of the nineteenth century in the period of La Grande Hydraulique. Some of these have already filled with sediments (Figure 8.13b). In Murcia, Spain, the wall of the Valdeinfierno reservoir was extended vertically, while in the Puentes reservoir immediately downstream (Figure 8.14a) a new wall has been constructed in front of the original. This construction of reservoirs has had a marked impact on the flow regime of the Segura River (Figure 8.14b), completely changing it from one dominated by peak flow in winter months, to one of peak (but lower) flows in summer months. These changes reflect not only the winter rains, but also the summer demand for irrigation. The summer discharges are now controlled ‘ecological’ flows required to maintain the ecological health of the river as it passes through the city of Murcia and beyond. As a consequence of these and other measures, the quality of the Segura has notably improved since 1972. Because of the seasonal characteristics of both the industry (fruit processing) and the flows, the smaller

Hydrology, River Regimes, and Sediment Yield

241

Fig. 8.12. A water cistern directly under the former Greek agora in Ptolomais, Libya. It receives water via an aqueduct from mountain springs 25 km to the east (photo: courtesy of Paul Fordham).

tributaries, such as the Mula, were heavily contaminated by stagnant pools in the summer months (Victoria Jumilla and Vicente Lopez 1986). These conditions impose special demands for integrated river management under the European Water Framework Directive. Establishing an appropriate balance between water storage, water abstraction, and the ecological needs

of channel systems is a major water resource management issue across the Mediterranean (Thornes and Rowntree 2006). The loss of water storage capacity to sedimentation is a particularly important water resource management issue in Mediterranean northwest Africa. In Morocco, the average annual water storage loss to sedimentation in the thirty-four largest

242

John Thornes et al.

Fig. 8.13. (a) A newly built check dam in the Segura River basin, south-east Spain. The water and sediment will fill the area to the left of the main wall. The apron and raised hump to the right are to prevent a scour pit developing during flood events and undermining of the main wall. (b) A sediment-filled reservoir: Valdeinfierno Reservoir, Murcia, Spain. The water storage capacity is now less than 5 per cent of the original. The foreground is wave-rippled sediment and not water. The location of this reservoir is shown at Figure 8.14a (photos: John Thornes). (c) Oblique air photograph of the town of Puerto Lumbreras and the Rambla Nogalte (see Figure 8.14a) in the aftermath of the large flood in October 1973 (http://www.puerto-lumbreras.com/galeria.asp?galeria=3).

Hydrology, River Regimes, and Sediment Yield

(a) Rio

Cenajo Reservoir

Segura

243

Barcelona Madrid

Cenajo

SPAIN Almadenes Malaga Archena

Ri o

Segura

Murcia

tín

en

l da

ua

G

N

Rio

Valdeinfierno Reservoir Rambla de Torrealvilla Lorca

Puentes Reservoir

Cartagena

Puerto Lumbreras

00

Rambla Nogalte

m

10

0

25 km

40

(b)

-1

Discharge (m s )

30

3

Before 1932

After 1957

Cenajo Almadenes Archena

20

10

0 J

F

M

A

M

J

J

A

S

O

N

D

Fig. 8.14. (a) Map of the Segura River basin (18,800 km2 ) in south-east Spain, showing the location of the sites shown in Figure 8.13. (b) The impact of reservoir construction on the monthly distribution of flows in the Segura River in Murcia, south-east Spain. The three gauging stations are shown in (a) above.

244

John Thornes et al.

reservoirs amounts to about 0.5 per cent (Lahlou 1988). Reservoirs in Algeria are estimated to be losing 2–3 per cent of their storage capacity each year and this equates to about 90 million m3 of water (Grenon and Batisse 1989).

Runoff Generation at the Local Scale In the previous section we looked at how the catchment runoff of Mediterranean rivers has been both indirectly and directly modified over the centuries by human activities. In this section the main natural controls of runoff generation at the local scale are considered. With the coming of remote sensing, hydrological studies of Mediterranean catchments included modelling runoff on the basis of the distribution of rainfall over the catchment and its distribution of intensity through time. It became recognized that soil conditions controlled infiltration rates and hence runoff, especially where vegetation was very scarce. A growing body of research indicated that surface crusting was also important. This crusting might be compacted soils a few millimetres thick formed under heavy rain, or chemical precipitation such as Mediterranean calcretes (Chapter 6). The crust can also be formed by lichens or other micro-organisms. These various crusts seriously impede infiltration rates and produce more excess runoff. In large catchments, these variations in surface texture and therefore runoff can be enough to control the volume and timing of runoff and therefore of stream hydrographs. Even the simple discrimination between bedrock runoff and alluvial surface runoff can be enough to improve runoff predictions from dryland catchments greatly. The term ‘hydrologically similar surfaces’ (HYSS) is used to describe surface properties that affect runoff generation. A study of the 1,400 km2 Nahael Zin catchment in southern Israel by Lange et al. (1999), using remote sensing to identify HYSS and GIS techniques to map and model runoff, removed much of the uncertainty in runoff prediction. Field data were used to specify the characteristics of the HYSS. By way of clarification of this emerging concept, Bull et al. (2000) used an improved infiltration model coupled with digital runoff simulation to examine the major controls on runoff generation at different scales. The study was based on observations of the Nogalte and Torreavilla catchments in south-east Spain. At the hillslope scales, runoff intensity was recognized by classifying the morphological evidence that ranges

from ‘no evidence of flow’ to ‘streams and gullies >1 m deep’. The different morphological runoff types are arranged in different zones on different lithologies according to the length-slope product. For the simulated storms, the storm structure (cell size and intensity) was not important in producing at-a-point runoff even over a simulated catchment of 1,000 km2 . However, at these larger scales, the basin morphology has an important control on runoff generation, because it determines water storage on the hillslope and transmission losses into the channel bed. Also at this scale, the connectivity of the channel network becomes very important. In summary, the most important control on runoff for these simulated conditions is variations in intensity through a storm. For observed data on catchment characteristics (e.g. shape and connectivity) and storm cell size, the effect of storm size did not lead to large errors, if estimates of catchment runoff were made from a single central rain gauge. Large divergence only appeared for areas of 1,000 km2 or more. The final factor of great importance is the areal distribution of crusting properties as indicated above (Chirino et al. 2006). Runoff generation in Mediterranean catchment systems is rather different from that of weakly-seasonal temperate environments, as discussed by Beven (2002). In this important contribution, Beven described our perceptions of what constitutes the behaviour of ephemeral channels: the complexity of the controlling processes, the strong non-linearity of that behaviour, the extreme heterogeneity of the runoff surfaces, and the limited storm extents in space and time (Figure 8.15 Chapter 18). He concluded (p. 62) that ‘We still have much to learn about runoff generation in semi-arid catchments.’ In particular he questioned the assumption that flow in these catchments was dominantly Hortonian overland flow (when rainfall intensities are significantly higher than infiltration capacities). Kirkby (1969) showed that in temperate environments this was rarely the case and supported Dunne and Black’s thesis that hydrological modelling must accommodate sub-surface flow, especially where flow is convergent (Dunne and Black 1970). Beven (2002) went on to review evidence for subsurface flow even in dryland systems, reaching the conclusion that, in Mediterranean catchments, this can be readily accepted as occurring under winter conditions. The main problems are choosing appropriate infiltration models and capturing the huge heterogeneity at different scales, even down to the level of the individual plant (Figure 8.9). This seems to require an alternative approach to fully distributed hydrological models or

Hydrology, River Regimes, and Sediment Yield

245

Fig. 8.15. The Río Aguas at Urra in the Sorbas basin, Almeria, south-east Spain. (a) The river in flood with turbulent flows and high suspended sediment concentrations (photo: courtesy of Joe Walsh) and (b) when the channel is dry in April 2006 (photo: Jamie Woodward).

a better management of the parameterization problem. An example of the former is the stochastic approach to modelling, as in the work in the Rambla de Algeciras (Murcia) by Conesa-Garcia and Alonso-Sarria (1997) or by Shannon et al. (2002) which proposes a Markov routing approach to ephemeral channels. Beven’s chapter concluded with the second approach—improving the parameterization. This may seem a far cry from Mediterranean rivers, but the continued demand for water and buildings in the coastal zone where the rivers and ramblas (ephemeral gravel-floored channels) meet the sea means that flash floods there can have disastrous effects on property, so that effective identification of flood-prone areas is a very high priority (Chapter 18). Although this can in part be achieved by historical evidence and experimental hardware models (e.g. for groundwater recharge), mathematical hydrological modelling also offers a route that can be pursued.

Catchment Sediment Yields The steep relief of Mediterranean river basins, the widespread occurrence of erodible rock types, and the heavy storm events that can generate rapid surface runoff combine to produce high suspended and bed sediment loads across the region. The erosion processes responsible have been reviewed in detail in Chapter 6. While there is little doubt that human activity is also important—and bare fields with limited soil conservation measures can lose large volumes of top soil during storm events (Figure 8.10b)—the underlying topographical, lithological, and climatic controls combine to produce an environment with much higher sediment fluxes per unit area than more temperate areas in central

and northern Europe for example (Woodward 1995; Milliman and Syvitski 1992).

Suspended Sediment Yield Dedkov and Moszherin (1992) have attempted to distinguish between the anthropogenic and natural or background sediment yield of river basins in mountain environments across the world. This is a very difficult task and the database for mountain river basins is very patchy indeed. However, they used land use type and extent to classify catchments into natural or modified states. Many would dispute this finding, but their analysis suggests that the sediment loads of mountain river systems in the Mediterranean are influenced by human action to a greater extent than those in any other climatic zone (Figure 8.16). It is, however, difficult to generalize about such a diverse region when there are so many variables to consider. Nonetheless, what does seem clear is that within the Mediterranean region, clear contrasts are evident in the magnitude of suspended sediment yields from catchments in the north and south with the latter typically an order of magnitude higher that the former (Milliman and Syvitski 1992; Woodward 1995). Table 8.1 contains data for two large river catchments in the northern Mediterranean that drain into the Aegean Sea to the west of Thessaloniki in north-east Greece. The Axios (23,747 km2 ) and Aliakmon (9,250 km2 ) rivers drain humid headwater catchments with elevations well over 2,000 m and have mean annual water fluxes of 5,030 and 2,292 million m3 per year respectively (Kapsimalis et al. 2005). These catchments—which include a range of land use types— have mean annual suspended sediment yields of 555 and 670 t km2 yr−1 , respectively, which are much lower than the typical yields for catchments in the Maghreb

246

John Thornes et al.

Suspended sediment yield (t km-2 yr-1)

2000 Anthropogenic component Natural component

1500

1000

500

0 l ria to ua Eq nna est va for Sa cal an e i op an Tr terr i ed M t s re Fo e t p r ep se St -de t i r m se Se -de i m Se e pe st p e ep tep or St t s d f s e re av Fo d-le oa Br a l ig Ta iva n bSu al ci la

G

(Lahlou 1988; Woodward 1995; Woodward and Foster 1997). In a review of the sediment yield of African rivers, Walling (1984) identified the narrow strip of upland catchments across Mediterranean north-west Africa as the highest yielding river basins on the continent. Suspended sediment yields from many of these catchments range from 1,000 to 5,000 t km−2 yr−1 . In an average year, the total annual suspended sediment load from river systems in the Maghreb that drain to the Mediterranean Sea is about 100 million tonnes (Probst and Amiotte Suchet 1992), and the flood events responsible are commonly characterized by very high suspended sediment concentrations. Network expansion by gully development in soft sediments can provide an important source of suspended sediment in Mediterranean catchments (Figure 8.17a). Thick, single-event accumulations of fine-grained sediments in channel and floodplain zones are indicative of high suspended sediment concentrations and loads (Figure 8.17b). High suspended sediment loads are a feature of many Mediterranean river basins but this is often strongly controlled by catchment lithology. Figure 8.18 shows how sediment yields in Moroccan catchments are markedly reduced where resistant rocks such as quartzites and granites are important. In contrast, where

Fig. 8.16. Suspended sediment yield from river basins in mountain environments in different climate and vegetation zones based on the analysis of Dedkov and Moszherin (1992).

TABLE 8.1. Water and sediment fluxes from the Axios and Aliakmon rivers that drain into the north-west Aegean Sea Axios Catchment area (km2 ) Maximum elevation (m) Mean annual rainfall (mm) Land use (%) Forest cover Uncultivated area Arable land Urban area Wetland Water fluxes (106 m3 ) Mean annual discharge Monthly maximum discharge Monthly minimum discharge Material fluxes (106 t yr−1 ) Suspended sediment load Dissolved load Bed load

23,747 2,800 650

Total load

16.9 5.9 75.4 1.7 0.1 5,030 8,800 1,545

Aliakmon 9,250 2,200 750 19.0 40.0 39.3 1.3 0.4 2,292 4,320 662

13.2 1.7 2.6

6.2 1.2 1.3

17.5

8.7

Source: Modified from Kapsimalis et al. 2005.

shales, marls, flysch, and other erodible lithologies are extensive, sediment yields can exceed 6000 t km2 yr−1 .

Hydrology, River Regimes, and Sediment Yield

247

Syvitski et al. (2005) estimated that human activity (mainly through land use change in river catchments leading to soil erosion) has increased the global sediment load of river systems by 2.3 ± 0.6 billion metric tons per year whilst simultaneously reducing the flux of sediment reaching the marine environment by 1.4 ± 0.3 billion metric tons per year due to sediment retention in reservoirs. Dam construction for water storage has been a favoured response to the pronounced seasonality in Mediterranean river flows since Roman times and many of today’s catchments contain impoundments at a range of scales. In the case of river systems draining to the Mediterranean and Black Seas, Syvitski et al. (2005) have estimated that 30 per cent of the fluvial suspended sediment flux is retained behind reservoirs and this compares to a global average of 20 per cent. Interestingly, their analysis also shows that Mediterranean river systems have the lowest proportion of summer (JJA) and autumn (SON) sediment fluxes of any landmass or ocean basin at 9 and 7 per cent respectively of the annual total.

Bed Load Fluxes

Fig. 8.17. (a) Gully erosion in soft sediments in central Israel (photo: Ian Foster) (b) A veneer of fresh suspended sediments deposited within the channel zone of the Torcicoda River in central Sicily (photo: Jamie Woodward).

River monitoring studies and reservoir sediment surveys have shown that river catchments in the Maghreb region of North Africa have some of the highest suspended sediment yields in the circum-Mediterranean zone (Woodward 1995; Woodward and Foster 1997; McNeill 1992). Water storage losses to sedimentation are a major concern for water resource managers in Morocco, Algeria, and Tunisia (Chapter 21). In a recent global-scale analysis of human impacts on the flux of sediment from the land surface to the oceans,

Information on bed load transport in Mediterranean river systems is generally much patchier than the database on suspended sediment transport. However, valuable datasets are available from Israel in particular where pioneering work on the ephemeral rivers of the Negev has provided important insights into the dynamics of the region’s dryland rivers (e.g. Laronne and Reid 1993; Reid and Laronne 1995; Reid et al. 1998). Bed load transport is difficult to measure and it poses particular logistical problems in semi-arid and arid river catchments because floods are so infrequent and unpredictable. In this context, it is useful to consider some of the hydrological characteristics of dryland catchments in the Mediterranean. A well-researched example is the Nahal Eshtemoa catchment (119 km2 ) that drains part of the southern Hebron Hills and is typical of the river systems of the northern Negev Desert. This catchment receives an annual rainfall input of 220 to 350 mm within a region where the potential evaporation is commonly >2,000 mm per year. This produces large soil moisture deficits which, in combination with sparse vegetation cover and low infiltration capacities, can produce overland flow within minutes of the onset of rainfall (Reid et al. 1998). Flow duration analysis of the Nahal Eshtemoa catchment for the period 1991–5 revealed that the channel conveyed flows on only seven days per year (2% of the time) and that overbank flooding took place for only 0.03 per cent of the time (Reid et al. 1998). The number of flash floods over this period

248

John Thornes et al.

Fig. 8.18. Suspended sediment yield from Moroccan river catchments formed in different rock types (modified from Lahlou 1988).

ranged from two to seven per year and most were produced by convective storms. The trunk stream hydrographs for these events show very spiky behaviour with most of the floods showing very steep rising limbs and multi-peaked hydrographs. Due to the roughness of the coarse-grained channel bed, flash-flood bore velocities are typically quite slow falling within the range 1–1.5 m per second (ibid.). Such infrequent and unpredictable flow regimes are typical of the eastern and southern Mediterranean and contrast markedly with the more humid regions in the north (Figure 8.7). The database on bed load flux in the ephemeral river systems of the Negev shows very high rates of sediment flux in comparison to perennial rivers of more humid environments. Laronne and Reid (1993) have argued that bed load data for the Nahal Yatir catchment show it to be 400 times more efficient at transporting coarse sediment than some of its perennial counterparts in more humid zones. Annual bed load sediment flux in the neighbouring Nahal Eshtemoa system was calculated as 4,641 tonnes or 39 tonnes km−2 while suspended sediment yield over the same period was 433 tonnes km−2 yr−1 giving a bed load to suspended load ratio of 1:11 (Reid et al. 1998). Data on bed load transport are now also available for perennial river systems in the mountains of the Mediterranean basin. For example, Garcia et al. (2000) employed an automated bed load monitoring system in a small catchment (35 km2 ) in the upper reaches of the Todera River in the coastal mountains of Catalonia in north-east Spain and monitored five flood events between 1995 and 1996. A key observation was the marked variability in bed load flux during the course of each flood event. Distinctive pulses of bed load transport were associated with sudden increases in flow depth and bed load response to hydraulic variables such as critical shear stress varied markedly between flood events. Much further downstream in the

lower, ephemeral, part of the same river basin, Rovira et al. (2005) assembled a fluvial sediment budget for an 11-km reach that lies just upstream of the basin outlet (894 km2 ) to the Mediterranean Sea. For the period from January 1997 to June 1999 they showed that the bulk of the total sediment load was transported as coarse bed load sediment. Around 156,000 tonnes of sediment (80% as bed load and 20% as suspended load) entered the study reach and approximately 107,000 tonnes (83% as bed load and 17% as suspended load) of sediment exited the reach and were discharged to the coastal zone and the Mediterranean Sea. Using repeat field surveys of the channel and floodplain zone in association with monitored sediment flux data, Rovira et al. (2005) estimated that 49,600 tonnes of sediment were deposited within their study reach over this period giving a mean rate of sediment accumulation on the channel bed of 6.8 mm per year. The extraction of bed load materials for aggregates and other commercial uses is widespread in the Mediterranean and this activity can result in dramatic modifications to coarse sediment budgets in river channel systems (Nicholas et al. 1999). Information on bed load transport should form part of the design of catchment management plans to ensure that such resource exploitation is sustainable in the long term. This aspect of direct river channel modification is discussed in more detail in Chapter 11.

River Catchment Management and Hazard Mitigation Floods and Flood Protection Damage and loss of life from floods is commonplace in Mediterranean river catchments (Chapters 3 and 18) and appears to have been so throughout history. For south-east Spain, Gil Olcina (1971) records the floods of

Hydrology, River Regimes, and Sediment Yield

Lorca (Murcia) in the seventeenth century and Pocklington (1986) describes flooding of the huerta (cultivated area) of Murcia in Arabic times. Chapter 11 documents flood histories for the Little Ice Age in various parts of the Mediterranean including southern France, Corisca, Crete, and north-west Greece. An example from recent times was the flash flood of Puerto Lumbreras (Figure 8.13c). On 18/19 October 1973 an intense storm crossed southern Spain, from Torremolinos in the west to Benidorm in the east. Over 250 mm of rain fell on the mountains of the Sierra Nevada in the headwaters of rivers that drain to the Mediterranean coast, the Guadalhorce, Guadalfeo, Andarax, and Almanzora. The storm created immense damage to roads, bridges, irrigation installations, and agricultural land. It caused severe loss of life in the village of La Rabita, where homes were destroyed by a mudflow emerging from a small canyon. In the town of Puerto Lumbreras (Figure 8.13c), where the weekly market was being set up in the dry river-bed, the flood wave washed away the entire market. Houses and cave dwellings and Moorish irrigation canals were destroyed by the on-rush of water. Today the scene is more tranquil. At Puerto Lumbreras, huge walls have been constructed on either side of the channel to protect the remaining town. The catchment of the Río Nogalte (Figure 8.14a) has been contour ploughed and trees have been planted. Over a hundred check dams have been installed in the steep-sided tributaries (barrancos) by the government agency responsible for natural resources. The Ministry of Public Works has reconstructed roads and bridges. Houses have been repaired, sub-alluvial galleries reinstated and the market continues to operate in the river-bed on a regular basis. The more protection that is provided, the more risk the population appears to be prepared to take. Although statistics suggest the event has a probability of occurring about 1 in 500 years, an equally disastrous storm occurred in 1982 in the Valencia area, immediately to the north. This story is repeated frequently across the Mediterranean and reflects a combination of intense storms, poor vegetation cover, and a long period of land management that encourages high and rapid runoff from mountainous terrains and impervious soils. The storm and flood hazards in the Mediterranean region are discussed in greater detail in Chapter 18.

River Catchment Management In the EU MEDALUS (Mediterranean Desertification and Land Use) Project, the role of flooding in the Guadalentin River basin (Figure 8.14a) was recognized

249

as a component of the problem of desertification and a plan was devised to manage the problem through land use control in the catchment of one of its tributaries— the Rambla Nogalte (Rojo Serrano et al. 2002). This was already partly underway in the remedial work following the 1973 flood. The rainfall of the basin ranges from 200 mm per year in the lowland areas, to 800 mm in the mountain headwaters and the rainfall is characterized by high intensities and short duration. Half of the basin has slopes greater than 12 per cent (at which erosion starts) and about one-third has slopes greater than 20 per cent (the overland flow-rilling threshold). A very small area of the basin is covered with sclerophyllous forest in the mountains. At lower altitudes and over most of the basin, the conditions are too dry for oak woodland (Quercus rotundifolia), but pines (Pinus pineaster, nigra, and halepensis) can survive. At even lower altitudes, bushes of the matorral prevail. Overall, 45 per cent of the basin is covered with ‘wildland’ and only 25 per cent of the basin is covered by forest, with another 20 per cent by shrub and bush communities. Water scarcity, groundwater exploitation, and flood protection are the most important environmental issues, followed by soil erosion. In the past, mechanized afforestation techniques in this catchment have been more effective than manual ones in decreasing hillslope runoff, retaining and storing as much water and moisture as possible (Rojo Serrano et al. 2002: 310). Pinus halepensis is the fundamental afforestation species. It is a Mediterranean species and its resistance to excessive sunshine means it has no need for shade when first planted. However, in the worst climatic and edaphic situations, this species is unable to reach forest status. In this study, specific management techniques were recommended for the Upper Guadalentin, ranging from better management of the existing forest and grazing areas, to restoration of saline soils in the lowland areas. This example provides a model approach that could be applied in many flood-prone areas of the Mediterranean and beyond. In other management schemes, unexpected effects may lead to unexpected and undesirable results. Borel (1994) showed how the construction of the SerrePonçon dam in the headwaters of the River Durance in Provence had entirely changed the ecosystem of this steep gradient river. The barrage was completed in 1959. The pre-impoundment river experienced enormous seasonal variations, with very low flows in the summer droughts to devastating floods (6,000 m3 /s−1 ). The Serre-Ponçon dam—with its 1270 million m3 of storage—now completely regulates the flow, but it is only one of a series of dams on the Durance. The

250

John Thornes et al.

of north and north-west Europe and even there they raise three key questions. First, which nature should it be? Many states of nature can be created—prehistoric, pre-industrial, or even pre-glacial (Hull and Robertson 2000). Second, can universal indicators be devised that are suitable for most aspects of river restoration in different environments, cultures, and socio-economic constraints? Third, can the rules, values, objectives, and procedures developed for the temperate rivers of northwest Europe be sensibly applied to the seasonally controlled humid and ephemeral rivers of the European Mediterranean?

flow is now more regular, with a diminished snowmelt flood in spring. Several new biotopes have appeared in the channel and floodplain complex and a significant increase in riparian vegetation, especially reed beds, has resulted in an increase in animal and plant diversity in the river corridor. The modified flow regime has also created a new corridor for the spread of plant and animal species. The changes in biodiversity highlight the need for good management and careful attention to the impact of changes on river ecosystems. This example also highlights a common problem throughout the Mediterranean region—the tension between what is desirable for mountain catchment areas and the resource needs of the littoral plains where economic activities and population are concentrated (Chapter 21).

Concluding Comments: Problems and Prospects

The Water Framework Directive

In the seasonal rivers of the Mediterranean there are quite different management priorities from those of temperate Europe. These include:

In December 2000 the European Union published the Water Framework Directive (Directive 2000/60/CE). This is a legal requirement on all member states to create an information base and an implementation plan so that all inland and coastal waters will reach ‘good status’ by 2015. They will do this by establishing a river basin district structure within which demanding environmental targets will be set, including ecological measures, for surface waters. Monitoring networks for water quantity and quality are patchy across the Mediterranean region and this represents a very significant management problem. The paradigm is now established and we briefly consider here whether it provides an appropriate model for Mediterranean rivers, either those already within the Union, directly or by association, or for those outside it. This Directive reflects the emergence of three important paradigms of river management, worldwide. First, rather than seeking to continue the modification of rivers that has taken place throughout the last two millennia, river managers should seek to restore rivers as close as possible to ‘natural’ conditions (Brookes and Shields 2001). Second, to do this, the river must be seen as a product of its catchment and integrated catchment management (ICM) must be adopted. Consequently changes may be required in the catchment in order to restore the river. Third, this is part of an overall paradigm towards restoring nature that has come to the forefront of the environmental movement (Gobster and Hull 2000). The Water Framework Directive carries with it extensive documentation, including specific guidelines about how to proceed with the implementation. These are more easily implemented for the temperate river basins

r Many rivers have high sediment yields, so the focus

of interest in ICM is the erosion and transport of particulate matter in the catchment. Interest therefore has to focus as much on hillslope as channel process and more on semi-natural vegetation and agricultural land use than on recreation and forestry. r The seasonality of flow and the existence of large flood events shift the thrust of the ecological argument as much to the channel beds and the flood corridors as well as to the water quality issues. r Flooding is a high priority and consequently so is flood protection. The flashiness of Mediterranean rivers is well documented and deserves a high profile in restoration works. Insensitive elimination of water control structures in an overenthusiastic response to catchment management could destroy the work of centuries with disastrous outcomes. r In several important examples in the Mediterranean, irrigation requirements are coupled to inter-basin transfers by the question ‘What is the catchment?’ Different qualities of water may become mixed and in the recent lively debate in Spain over the proposed water transfer from the Ebro to the Segura catchments, it has sometimes been forgotten that the Ebro waters are naturally saline at the proposed take-off point as a result of the contribution from the highly saline marls and evaporite rocks of the middle and lower part of the basin just east of Zaragoza. The Segura River in Murcia illustrates all these issues. It has been transformed by human activity for over two

Hydrology, River Regimes, and Sediment Yield

millennia, from changes of land use in the Bronze Age to the recent water quality implications of the fruitprocessing industry in the towns along its course, that have resulted in a seriously polluted river. Many of the structural changes to its course were devised as a compromise between evacuating flood waters as fast as possible and encouraging aquifer recharge as much as possible. Significant areas of the catchment are severely eroded and there already exists a management plan to reduce erosion and its associated high runoff rates and sediment yield. The flooding problem is an important issue and dam construction is a major element in runoff control and flood protection. The Water Framework Directive represents a crucial shift in thinking towards the principles of integrated catchment management. However, the practice of integrated catchment management is broader than the practice of integrated water management which has been a keystone for many years in the control of Mediterranean rivers. Guidelines cannot be too prescriptive, but must take account of the diversity of both environment and culture in Europe. At the same time they must be rigid enough to ensure compliance. The implementation of the Water Framework Directive is a very costly and demanding activity to impose on those catchments that are thinly populated or in economically marginal areas. In closing this chapter we reiterate the point that it is impossible to understand the hydrology of Mediterranean catchments without appreciating the huge impact of human activities. On the other hand, nature is a strong force that cannot be ignored. Incursions into areas with a high flood risk, careless disregard for modifications of land use and exploitative use of groundwater all come at a high cost. Whilst it can be argued that the integrated approach of the Water Framework Directive in the dryland catchments of southern Europe would be inappropriate (Thornes and Rowntree 2006), it might yet yield important lessons for the integrated management of river catchments across the entire Mediterranean region.

Acknowledgements The authors thank the external reviewer who provided valuable comments on an earlier draft of this chapter. Nick Scarle in the Cartographic Unit at The University of Manchester expertly redrafted the figures. We thank Ian Foster for providing photographs of Israel and Tunisia.

References Beven, K. (2002), Run-off generation in semi-arid areas, in L. J. Bull and M. J. Kirkby (eds.), Dryland Rivers: Hydrology

251

and Geomorphology of Semi-Arid Channels. John Wiley & Sons, Chichester, 57–105. Bochet, E., Poesen, J., and Rubio, J. L. (2006), Runoff and soil loss under individual plants of a semi-arid Mediterranean shrubland: influence of plant morphology and rainfall intensity. Earth Surface Processes and Landforms, 31/5: 536–50. Borel, L. (1994), Influence des aménagements sur l’évolution des Milieux Durnciens—dynamique des peuplements vegetaux et animaux, in J. Riser (ed.), Aménagements et gestion des grandes rivières mediterranéennes. Étude Vauclusiennes 5: 15–19. Brookes, A. and Shields, F. D. (2001), River Restoration. John Wiley & Sons, Chichester. Bull, L. J., Kirkby, M. J., Shannon, J., and Hooke, J. M. (2000), The impact of storms on floods in ephemeral channels in south-east Spain. Catena 38/3: 191–210. Butcher, G. C. and Thornes, J. B. (1978), Spatial variability in runoff processes in an ephemeral channel. Zeitschrift für Geomorphologie Suppl. 29: 83–92. Carson, M. A. and Kirkby, M. J. (1972), Hillslope Form and Process. Cambridge University Press, Cambridge. Chirino, E., Bonet, A., Bellot, J., and Sanchez, J. R. (2006), Effects of 30 year old Allepo pine plantations on runoff, soil erosion and plant diversity in a semi-arid landscape in south-eastern Spain. Catena 31/1: 19–30. Christopoulos, G. (1998), Late Holocene river behaviour in the lower Alfios basin, Western Greece. Ph.D. Thesis, University of Leeds. Conesa-Garcia, C. and Alonso-Sarria, F. (1997), Stochastic matrices applied to the probabilistic analysis of runoff events in a semi-arid stream. Hydrological Processes 11: 297–310. Cudennec, C., Leduc, C., and Koutsoyiannis, D. (2007), Dryland hydrology in Mediterranean regions—a review. Hydrological Sciences Journal 52: 1077–87. Dedkov, A. P. and Moszherin, V. T. (1992), Erosion and sediment yield in mountain areas of the world, in D. E. Walling, T. R. Davies, and B. Hasholt (eds.), Erosion, Debris Flows and Environment in Mountain Regions of the World. IAHS Publication 209: 29–36. Dunne, T. and Black R. D. (1970), Partial area contributions to storm runoff-producing zones in a small New England watershed. Water Resources Research 6: 1296–311. Eagleson, P. S. (1978), Climate, soil and vegetation. 6. Dynamics of the annual water balance. Water Resources Research 14/5: 749–64. Elwell, H. A. and Stocking, M. A. (1976), Vegetation cover to estimate soil erosion hazard in Rhodesia. Geoderma 15: 61–70. Evenari, M., Shanan, L., and Tadmor, N. (1982), The Negev: The Challenge of a Desert. Harvard University Press, Cambridge, Mass. Francis, C. F. and Thornes, J. B. (1990), Runoff hydrographs from the Mediterranean vegetation cover types, in: J. B. Thornes (ed.), Vegetation and Erosion, John Wiley and Sons, Chichester, 313–84. (1994), Matorral: erosion and reclamation, in J. Albaladejo, M. A. Stocking, and E. Diaz (eds.), Soil Degradation and Rehabilitation in Mediterranean Environmental Conditions. Consejo Superior de Investigaciónes Científicas, Madrid, 87–116. Garcia, C., Laronne, J. B., and Sala, M. (2000), Continuous monitoring of bedload flux in a mountain gravel-bed river. Geomorphology 34: 23–31.

252

John Thornes et al.

Garcia-Ortiz, E. (2006), Efecto de la structura de la copa en la partición de lluvia de tres species arbustivas en clima semiarido. Doctoral thesis, University of Almeria. Gil Olcina, A. (1971), El Campo de Lorca: estudio de geografía agraria. CSIC, Valencia. Gobster, P. H. and. Hull, R. B (eds.) (2000), Restoring Nature: Perspectives from the Social Sciences and Humanities. Island, Washington. Gómez-Espin, J. M. (2004), Approvechamiente Integral del Agua en la Rambla de Nogalte (Puerta Lumbreras-Murcia), University of Murcia, Murcia. Grenon, M. and Batisse, M. (eds.) (1989), Futures for the Mediterranean Basin: The Blue Plan. Oxford University Press, Oxford. Grove, A. T. and Rackham, O. (2003), The Nature of Mediterranean Europe: An Ecological History. Yale University Press, New Haven. Hull, R. B. and Robertson, D. P. (2000), The language of nature matters: we need more Public Ecology, In P. H. Gobster and R. B. Hull (eds.), Restoring Nature: Perspectives from the Social Sciences and Humanities. Island, Washington. Imeson, A. C., Verstraten, E. J., Mulligan, J. M., and Sevink, J. (1992), The effects of fire and water repellency on infiltration and runoff under Mediterranean type forest. Catena 19: 345–61. Inbar, M. (1992), Rates of fluvial erosion in basins with a Mediterranean type climate. Catena 19: 393–409. Laronne, J. B. and Reid, I. (1993), Very high rates of bedload sediment transport by ephemeral desert rivers. Nature 366: 148–50. Kapsimalis, V., Poulos, S. E., Karageorgis, A. P., Pavlakis, P., and Collins, M. (2005), Recent evolution of a Mediterranean deltaic coastal zone: human impacts on the Inner Thermaikos Gulf, NW Aegean Sea. Journal of the Geological Society, London 162: 879–908. Kirkby, M. J. (1969), Infiltration, throughflow and overland flow, In R. J. Chorley (ed.), Water, Earth and Man. Methuen, London, 215–28. Bracken, L., and Reaney, S. (2002), The influence of land use, soils and topography on the delivery of hillslope runoff to channels in SE Spain. Earth Surface Processes and Landforms 27: 1459–73. Lahlou, A. (1988), The silting of Moroccan dams, in M. P. Bordas and D. E. Walling (eds.), Sediment Budgets. IAHS Publication 174: 71–7. Lange, J., Leibundgut, C., Greenbaum, N., and Schick, A. P. (1999), A non-calibrated rainfall-runoff model for large arid catchments. Water Resources Research 35: 2161–72. Lewin, J., Macklin, M. G., and Woodward, J. C. (eds.) (1995), Mediterranean Quaternary River Environments. Balkema, Rotterdam. Macklin, M.G., Lewin, J., and Woodward, J. C. (1995), Quaternary fluvial systems in the Mediterranean basin, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 1–25. McNeill, J. R. (1992), The Mountains of the Mediterranean World. Cambridge University Press, Cambridge. Mavromati, E. and Chryssaidis, L. (2007), Aqueducts in the Hellenic area during the Roman Period. Water Science and Technology: Water Supply 7: 139–45. Nicholas, A. P., Woodward, J. C., Christopoulos, G., and Macklin, M. G. (1999), Modelling and monitoring the impact of dam construction and gravel extraction on rates of bank

erosion in the Alfios River, Peloponnese, western Greece, in A. G. Brown and T. A. Quine (eds.) Fluvial Processes and Environmental Change. John Wiley & Sons, Chichester, 117–37. Obando, J. (2002), The impact of land abandonment on regeneration of semi-natural vegetation. A case study from the Guadalentin, in J. A. Geeson, C. J. Brandt, and J. B. Thornes, Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 269–77. Pocklington, R. (1986), Acequias árabes y pre-árabes en Murcia y Lorca: apportación toponímica a la historia del regadío, in X Colloqui General de la Societat d’Onomàstica, 1985. University of Valencia, 462–73. Probst, J. L., and Amiotte Suchet, P. (1992), Fluvial suspended sediment transport and Mechanical erosion in the Maghreb (North Africa). Hydrological Sciences Journal 37: 621–37. Reid, I. and Laronne, J. B. (1995), Bedload sediment transport rates in an ephemeral stream and a comparison with seasonal and perennial counterparts. Water Resources Research 31: 773–81. and Powell, D. M. (1998), Flash-flood and bedload dynamics of desert gravel-bed streams. Hydrological Processes 12: 543–57. Rojo Serrano, L., Garcia Robredo, F., Martinez Artero, J. A., and Martinez Ruiz, A. (2002), Management Plan to combat desertification in the Guadalentin River Basin, in N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 303–19. Rovira, A., Batalla, R. J., and Sala, M. (2005), Fluvial sediment budget of a Mediterranean river: the lower Tordera (Catalan Coastal Ranges, NE Spain). Catena 60: 19–42. Shannon, J., Richardson, R., and Thornes, J. B. (2002), Modelling event-based fluxes in ephemeral streams in L. J. Bull and M. J. Kirkby (eds.), Dryland Rivers: Hydrology and Geomorphology of Semi-Arid Channels. John Wiley and Sons, Chichester, 129–72. Struglia, M. V., Mariotti, A., and Filograsso, A. (2004), River discharge into the Mediterranean Sea: climatology and aspects of the observed variability. Journal of Climate 17: 4740–51. Syvitski, J. P. M., Vörösmarty, C. J., Kettner, A. J., and Green, P. (2005), Impact of humans on the flux of terrestrial sediment to the global coastal ocean. Science 308: 376–80. Thornes, J. B. (1977), Channel changes in ephemeral streams: observations, problems and models, in K. J. Gregory (ed.), River Channel Changes. John Wiley & Sons, Chichester. (2007), Modelling soil erosion by grazing: recent developments and new approaches. Australian Geographical Research 45: 13–26. and Rowntree, K. (2006), Integrated catchment management in semi-arid environments in the context of the European Water Framework Directive. Land Degradation and Development 17: 255–264. Victoria Jumilla, F. and Vicente Lopez, E. (1986), La contaminación de las aguas en la region de Murcia. Ministerio de Obras Publicas, Murcia. Vita-Finzi, C. (1978), Archaeological Sites in their Setting. Thames & Hudson, London. Wainwright, J. (1996), Infiltration, runoff and erosion characteristics of agricultural land in extreme storm events, S. E. France. Catena 26: 27–47.

Hydrology, River Regimes, and Sediment Yield and Thornes, J. B. (2003), Environmental Issues in the Mediterranean: Processes and Perspectives from the Past and Present. Routledge, New York. Walling, D. E. (1984), The sediment yields of African rivers, in D. E. Walling, S. S. D. Foster, and P. Wurzel (eds.), Challenges in African Hydrology and Water Resources. IAHS Publication 144: 265–83. Woodward, J. C. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water

253

Quality in River Catchments. John Wiley & Sons, Chichester, 365–89. and Foster, I. D. L. (1997), Erosion and suspended sediment transfer in river catchments: environmental controls, processes and problems. Geography 82/4: 353–76. Macklin, G., Krom, M. D., and Williams, M. A. J. (2007), The Nile: evolution, Quaternary river environments and material fluxes, in A. Gupta (ed.), Large Rivers: Geomorphology and Management. John Wiley & Sons, Chichester, 261–92.

This chapter should be cited as follows Thornes, J. B., López-Bermúdez, F., and Woodward, J. C. (2009), Hydrology, river regimes, and sediment yield, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 229–253.

This page intentionally left blank

9

Lakes, Wetlands, and Holocene Environmental Change Neil Roberts and Jane Reed

Introduction The Mediterranean regions of the world are defined on the basis of their climate, with a distinct hot, dry summer season and a warm, wet winter (Grove and Rackham 2001; Chapter 3). Spring and autumn seasons are less well defined but often contribute significantly to annual precipitation. Strictly defined in this way, the Mediterranean region is confined to parts of Italy, Greece, southern France, the south and east of Spain (non-Atlantic climate), the Maghreb and Cyrenaica in North Africa, and narrow coastal strips running through the Balkans, southern and western Turkey, and the Levant (Syria, Lebanon, and Israel-Palestine) (Figure 9.1a). Outside these areas, climate becomes humid temperate (western Europe, Black Sea), arid (Sahara, northern Arabia), or continental (interior areas of the Balkans, Turkey and Iberia, the Zagros mountains of Iran/Iraq). Even within the strict definition are found subalpine mountain zones, so it is a difficult study region to demarcate absolutely. In a similar vein to the volume by Zolitschka et al. (2000), this chapter extends the scope to important wetlands in some neighbouring regions, and deals effectively with the circum-Mediterranean. Thus, we include lakes Ohrid and Dojran in the Balkans, wetlands of the continental interior of Turkey, north-western Iran and the Caucasus (e.g. Lakes Van, Urmia, and Sevan), the climatically dry Jordan rift valley which includes the Dead Sea, and the subalpine northern Italian lakes such as Como and Maggiore. The Mediterranean basin is geologically complex and has its origin in the progressive closure of the Sea of

Tethys during the Tertiary (Laubscher and Bernoulli 1977). Plate convergence between Africa and Eurasia led to a major phase of orogenesis and the creation of fold mountains including the Atlas, Sierra Nevada, Alps, Apennines, and Taurus, and to plateau uplift in Iberia and Anatolia (Chapter 1). These mountain ranges are commonly dominated by massively deformed Mesozoic limestones that now form karst landscapes (e.g. Dinaric Alps; Ager 1980; Chapter 10). Tectonic movement also led to extensive late Cenozoic volcanism, notably in southern and central Italy, the Hellenic arc, Anatolia, and around the Jordan rift (Chapter 15). The Alpine orogeny created not only mountain ranges, but also numerous intervening sedimentary basins, many of which are now occupied by lakes that can be large and sometimes deep (Chapter 4). This geological background is crucial to understanding lake genesis, and has led to important local and regional diversity in topography, rock structure and composition (and hence groundwater hydrology and surface water chemistry), and climatic conditions. Our definition of wetlands includes lakes, marshes, and coastal lagoons, with a focus on natural lakes. The high variability in Mediterranean wetland types has implications both in terms of strategies for conservation and for interpretation of their palaeolimnological records. Different lakes exhibit different thresholds of response to a given forcing function, and there is an unusually great need to understand each lake individually before moving on to attempt any regional syntheses of, for example, lakes as signals of regional climate change.

256

Neil Roberts and Jane Reed

(a)

(b) 12

8 7 9

3 11

2

10 6 5

1

4 0

1000 km

(d) (c)

(c) 20 19

21

17 18

13

27 15 16

24 14

23

25

28

29

30 31

26

32

22

Fig. 9.1. Maps showing: (a) Mediterranean type climates; (b) the location of the largest lakes in the circum-Mediterranean region >200 km2 in area (see Table 9.1 for details); (c) exemplar lake types (see Table 9.2 for details); (d) location of selected key Holocene palaeolimnological sites (see Table 9.3 for details).

Taxonomy and Distribution of Mediterranean Lakes and Wetlands The origins, geomorphology, hydro-chemistry, and biology of any lake, lagoon, or marsh are intimately interlinked with each other and with their geologic and climatic setting (Timms 1992; Wetzel 2001). Mirroring the topographic and geological complexity of the region, Mediterranean wetlands have varied origins when compared to many other ‘lake districts’ such as those of northern Europe and North America which are dominated more uniformly by lakes that are the legacy of Pleistocene glaciation. The location, origins, and main morphometric parameters of a range of Mediterranean lakes are described in Tables 9.1 and 9.2, along with some of the main human impacts upon them. They include many of the designated Ramsar sites (Ramsar 2004) and key sites monitored by the International Lake Environment Committee (ILEC; ). These lakes have been selected to represent the overall variety of different types in the circum-Mediterranean region, along with those that have been the subject of intensive palaeoenvironmental research (Table 9.3) or those that

are large and/or have high economic or ecological status (Table 9.4).

Genesis and Morphology Tectonic Many Mediterranean lakes and wetlands are tectonic in origin. One of the most remarkable is Lake Ohrid on the border between Albania and Macedonia (Former Yugoslav Republic of Macedonia: FYROM). This is Europe’s best example of a steep-sided, quadrangular graben, formed by parallel faulting and subsidence since the Late Miocene (Stankovi´c 1960). On a par with Baikal, Tanganyika, and Malawi, it is ranked amongst the most ancient lakes in the world, having been formed at least 4 million years ago. Other grabens include the nearby but smaller lakes of Prespa and Dojran, or the Greek lakes of Trichonis and Amvrakia. Most other Mediterranean ‘great lakes’ also owe their origin—at least in part—to major long-term tectonic faulting. There are good examples of fault-bounded lakes in Turkey including Sapanca and Iznik (north-west), Burdur, E˘girdir, and Bey¸sehir (south-west), and Hazar (south-east). The Jordan is the most famous of all rift valleys in the Mediterranean region and this contains the

TABLE 9.1 Individual characteristics of all permanent natural lakes in the circum-Mediterranean region >200 km2 in area, excluding coastal lagoons Name

Country

Genesis

Lat N Long E Altitude masl

Lake area km2

Catchment area km2

Max depth m

Salinity g/l

Notes

Sources

Ramsar, Holocene lake sediment record Sediments varved, Late Pleistocene–Holocene sediment cores Ramsar, switch from oligo–to meso-trophic, 19 m lake-level fall since 1933 Falling water levels, southern basin now isolated and used for salt pans, Pleistocene Lisan beds exposed around lake Significant commercial fisheries Former important crayfish production, now mostly lost Ramsar, largest bird colony in Europe, incl. pelicans Lake tourism, partial recovery from pollution World Heritage Site, oligotrophic

Schweizer 1975; Kelts and Shahrabi 1986 Kempe 1977; Degens et al. 1984, Wick et al. 2003

1. Urmia (Rezaiyeh) 2. Van

Iran

Tectonic

37 30 45 30

1280

4750

51000

13

95

Turkey

Lava-dammed

38 30 43 00

1648

3574

12522

451

21

3. Sevan

Armenia

Tectonic

40 20 45 20

1905

1360

3390

86

<1

4. Dead Sea

Israel/Jordan

Tectonic graben

31 40 35 40

−395

810

40650

330

300

5. Bey¸sehir ˘ 6. Egirdir

Turkey Turkey

Tectonic Tectonic

37 60 31 50 37 50 30 50

1123 920

651 590

4052 3309

10 14

<1 <1

7. Skadar

MontenegroAlbania Italy

Tectonic/karst

42 12 19 17

5

372

5490

8

<1

Glacial-tectonic

45 40 10 41

65

368

2260

350

<1

Macedonia/ Albania

Tectonic graben

41 05 20 45

695

348

1129

298

<1

10. Iznik

Turkey

Tectonic graben

40 43 29 53

85

308

nd

87

<1

11. Prespa

Macedonia/ Greece Italy/ Switzerland

Tectonic graben, karstic Glacial-tectonic

40 56 21 01

852

266

nd

55

<1

45 47

194

212

6387

370

<1

8. Garda 9. Ohrid

12. Maggiore/ Lugano

Note: The locations are shown in Figure 9.1b. (nd = no data).

8 40

Use for irrigation, receives untreated waste discharge Nature Reserve, Ramsar Lake tourism, water pollution but eutrophication reversed

World Lakes Database; www.unece.org/env/epr/ studies/armenia Neev and Emery 1967; Water Data Banks Project

Magnin and Yarar 1997 Magnin and Yarar 1997 World Lakes Database Premazzi et al. 2003 Allen and Ocevski 1976; Roelofs and Kilham 1983; Spirovski et al. 2001; Stankovic´ 1960 Magnin and Yarar 1997 Ramsar database Premazzi et al. 2003; World Lakes Database

TABLE 9.2 Exemplars of Mediterranean lake types Name

Country

Genesis

Lat N Long E

Altitude masl

Lake area km2 42.8

Catchment area km2

Salinity g/l

10.4

<1

13. Dojran

Macedonia/ Tectonic volcanic, 41 13 22 45 Greece karstic

147

14. Kinneret (Tiberias/ Galilee)

Israel

Tectonic graben

32 48 35 35

905

168

2730

43

<1

15. Tuz

Turkey

38 45 33 23

905

1900

nd

2

300

16. Nar 17. Banyoles

Turkey Spain

38 22 34 27 42 08 2 46

1363 175

0.56 1.12

2.4 11.42

18. Laguna de Pétrola

Spain

Tectonic plateau, playa Volcanic crater Karstic polje on fault line Tectonic: minor fault Dissolution

19. Laguna de los Spain Tollos 20. Como Italy 21. Lagunas de Ruidera Spain

22. Bardawil lagoon 23. Qarun, Faiyum

420

Max depth m

26 46.4

3.5 <1

Significant commercial fisheries, statistics pre-drainage; now down by 45% Monomictic, provides 30% of Israel’s water supply, commercial fisheries Receives waste water from Konya, flamingo breeding Stratified, varve formation Ramsar Classic playa lake

38 51

1 34W

860

1.74

nd

0.7

10

36 51

6 01W

70

0.7

nd

0.04

36

Subalpine, glacial 46 10 Fluvio-lacustrine 38 55 travertine dams

9 16 2 50W

198 840

145 1.8

4522 nd

Egypt (Sinai) Coastal lagoon Egypt Tectonic, aeolian

Note: The locations are shown in Figure 9.1c. (nd = no data).

31 05 33 05 29 30 30 30

Sea level 30 30

600 240

nd 1700 + Nile

410 2.7

<5 9

Notes

Formerly permanent; now ephemeral from drainage <1 Lake tourism, water pollution <1 National park; 15 linked lakes with high biodiversity; precipitating carbonate Hypersaline Ramsar, key winter wildfowl site 37 Fed mainly from Nile flood

Sources

Stankovic´ 1931; Griffiths et al. 2002; Zacharias et al. 2002 World Lakes Database; Hambright et al. 2000 Magnin and Yarar 1997 Jones et al. 2005 World Lakes Database Montes and Martino 1987; Reed 1995 Reed 1995 Premazzi et al. 2003 González Martín et al. 1987; Reed 1995; Nevado et al. 2004 Ramsar database http://www.ecrc.ucl.ac. uk/minisite/qarun/ index.htm

TABLE 9.3 Selected key Holocene palaeolimnological records for the Mediterranean Name

Lat N

Long E

Dissolution

36 37

6 03W

25. Tigalmamine Morocco

Karst

32 54

5 21

1630

0.06

3.5

16

26. Pergusa 27. Mezzano

Sicily, Italy Italy

Probable tectonic volcanic maar

37 31 14 18 42 37 11 56

667 452

1.4 0.5

7.5 1

4.6 31

28. Albano

Italy

Volcanic maar

41 30 12 40

293

6

3.68

29. Ioannina Greece (=Pamvotis)

Tectonic/karst

39 45 20 51

470

30

30. Eski Acıgöl

Turkey

Volcanic maar

38 33 34 33

1270

31. Zeribar

Iran

35 32 46 07

1300

32. Birket Ram

Syria-Israel

Tectonic-alluvial dam Volcanic maar

33 15 35 40

940

24. Medina

Country

Spain

Genesis

Note: The locations are shown in Figure 9.1d. (nd = no data).

Altitude masl 35

Lake area km2 1.2

Catchment area km2 16

Max depth m 0.7

Salinity g/l

Notes and length of available record

Sources

15.7

9 ka record; once permanent but now dry in drought years Clearwater, charophyte beds; >12 ka Sulphur bacteria blooms, ∼20 ka Meso- to oligotrophic, now has artificial outflow, 34 ka Stratified, varved, eutrophied since C19th; 30 ka Drainage, wastewater discharge, 430 ka

Reed 1995; Reed et al. 2001 Benkaddour 1993; Lamb and van der Kaars 1995 Sadori et al. 2001 Ramrath et al. 2000; Sadori et al. 2004 Guilizzoni and Oldfield 1996 Tzedakis this volume, Frogley et al. 2001; Frogley and Preece 2004 Roberts et al. 2001; Kazancı et al. 1995 van Zeist and Bottema 1977; Stevens et al. 2001 Ehrlich and Singer 1976; Schwab et al. 2004

<1 Up to 127 <1

175

<1

330

10

<1

0.4

1

na

6

nd

5

0.45

1.5

12

Now drained Late Pleistocene—Early Holocene varves, >16 ka <1 ∼40 ka <1

∼50 ka + 6.5 ka

TABLE 9.4 Selected key Mediterranean wetlands requiring conservation or restoration Name

Country

33. Doñana

Spain

34. Tablas de Daimiel

Genesis

Lat N

Long E

Altitude masl

Lake marsh area km2

Coastal/fluvial

36 53

6 25W Sea level

Spain

Karst/tectonic

39 09

3 40W

1800 (1900) 300 (1988) Reduced from 12 km long

35. Ichkeul

Tunisia

Karst/tectonic

37 10

9 40

36. Trasimeno

Italy

Volcanic

37. Burdur

Turkey

Tectonic graben

620

Catchment area km2

Max depth m

3000

na

19.3

na

Salinity g/l Variable

World Lakes Database

Magnin and Yarar 1997

78–127

2,080

Shallow

41 01 12 01

257

124

396

6.3

37 44 30 11

850

237 (1970) 182 (2002)

3874

75 (1970) 65 (2002)

38. Konya, including Turkey Hotami¸s, ˘ Eregli, Karapınar marshes 39. Küçükçekmece Turkey

Tectonic plateau, 37 30 33 30 some playas and karst

1000

20280

<2

Coastal lagoon

41 00 28 45

Sea level

15

nd

20

Mesosaline

40. Larnaca salt lake Cyprus

Coastal lagoon

34 52 33 33

1 to 10

16

nd

1

Hypersaline

41. Huleh

Tectonic graben, 33 04 35 37 volcanic dam

70

14 + 40 pre-drainage

∼2500

5 (pre-drainage)

Israel

Note: The locations shown in Figure 9.13b. (nd = no data).

Sources

Ramsar, key winter wildfowl site; eutrophication, drainage. Fresh + some Ramsar, key winter wildfowl site, playas much reduced by drainage and groundwater abstraction 10 to 40 Ramsar, key winter wildfowl site, stratified, varved, inflows diverted for irrigation, wastewater discharge <1 Significant commercial fisheries, serious eutrophication c. 30–40 Ramsar, key winter wildfowl site, stratified, varved, inflows diverted for irrigation, wastewater discharge Fresh + some Complex of marshes and shallow playas lakes now desiccated despite protected status, site of large Pleistocene lake

Sea level

∼700 (1980)

Notes

<1

Lake catchment now largely urbanized, extreme pollution Lake catchment now largely urbanized, extreme pollution Largely drained from 1950s, contains long sedimentary record

Bernués Sanz 1990a Bernués Sanz 1990b

Thomas et al. 1991; Ramsar database

Magnin and Yarar 1997; Roberts et al. unpublished

de Meester 1970; Roberts 1983; Magnin and Yarar 1997; Fontugne et al. 1999

Ramsar database Hambright and Zohary 1998

Lakes and Wetlands

Dead Sea, which is not a sea at all but a >300-m deep non-outlet salt lake. The Jordan rift forms the northern portion of the great fault system also occupied by the Red Sea and East African rift valleys to the south. Other tectonic lakes include Urmia (or Rezaiyeh, in north-west Iran), Sevan (Armenia), and Van (eastern Turkey). Lake Van, the largest lake by volume in the entire circumMediterranean region and the biggest soda lake in the world, has its origins in Pleistocene volcanic-tectonic damming of a tributary valley of the Tigris drainage system. At the other morphometric extreme, the uplifted plateaux of central Iberia and Anatolia have large expanses of relatively flat terrain where minor faulting or subsidence in regions of poor drainage has contributed in some cases to the formation of shallow depressions containing marshes and lakes which are often ephemeral and saline, such as the Laguna de Pétrola in Spain (Ordoñez et al. 1973), or the Konya basin wetlands of south-central Turkey (Magnin and Yarar 1997). Critical to the morphometry of all tectonic lakes (and the special case of dissolution lakes described below) is the balance between the rate of subsidence and the rate of sedimentary infilling. Where the former has exceeded the latter, space is created to accommodate a potentially deep water body (e.g. Ohrid, Burdur), but where the two have been broadly in balance, the overlying lake will be shallower, for example Lake Ioannina (or Pamvotis) in north-west Greece (Figure 9.2 and Chapter 4). In both cases, the resulting lake sediment record can be long and unbroken over tens to hundreds of thousands of years, as cores from Ioannina and Tenaghi Philippon in Greece and Castiglione in central Italy testify (Chapter 4). Where tectonic subsidence is slower than the infilling rate, by contrast, the basin will be ‘overfilled’ with sediment and is likely to support lakes that have been geologically ephemeral. In overfilled basins such as Konya, lake phases have alternated with periods of desiccation and erosion, and the Quaternary palaeolimnological record is characterized by sedimentary hiatuses (Roberts et al. 1999).

Karstic and Evaporitic Solution Basins Being dominated by limestone-rich Mesozoic sediments of the Tethys Sea, the Mediterranean is notable for the high number of lakes whose origins are related to karstic processes or dissolution of ancient gypsum/halite evaporites. The latter formed widely during the Late Miocene Messinian Salinity Crisis when the Mediterranean Sea desiccated (Chapter 1). Many of the numerous Spanish and North African ephemeral lakes (in North Africa,

261

large chotts and small, surface-fed sebkhas, and, in Spain, playa lakes) owe their origins in a large part to evaporite dissolution (e.g. Montes and Martino 1987; Florín et al. 1993 for Spain). More classic karst features are found in parts of Greece and the Balkans, southern Turkey, south-eastern Spain, and the Atlas Mountains of North Africa (Chapter 10), where the landscape is dotted with small, freshwater, circular lakes occupying conical dolines and other solution basins. The Tigalmamine system in Morocco’s Middle Atlas, for example, contains a cascade of three solution basins, the upper one (Admer) now infilled, but the lower two containing relatively deep pristine lakes (Benkaddour 1993). The Laguna de la Cruz in south-eastern Spain forms an almost perfect circular depression with steep walls rising above the water surface and is one of the few Spanish karstic lakes to lack gypsum in its catchment (Vicente and Miracle 1988). In some cases solution basins have become interconnected along fault lines to create larger polje basins such as Plitvice in Croatia (Stankovi´c 1960; Chapter 10). This combination of tectonic and karstic lake origin is common throughout the southern Balkans and Greece (Zacharias et al. 2002). Lago de Banyoles (Spain) is another example of a mixed tectonic-karstic lake, in this case comprising a series of six interconnected dolines (Moreno-Amich and GarcíaBerthou 1989). In some cases, bicarbonate-rich spring and river waters have built natural tufa barrages behind which small riverine lakes have been impounded, as in Spain’s Ruidera lakes on the upper Guadiana River, central Spain (González Martín et al. 1987).

Volcanic Maar and crater lakes are an important feature of Italy and Turkey. Most of the deeper lakes of the central Italian Peninsula are maars of volcanic origin, such as Albano, Vecchienna, Bolsena, Mezzano, Monticchio, and Vico. In central Turkey, lakes have also formed in ancient craters, for example Nar, Gölcük (Isparta), and around Nemrut Da˘g. One of the saline crater lakes near Karapınar on the Anatolian plateau has a volcanic cone island pushing up through its centre to create an unusual doughnut-shaped lake (Figure 9.3). Volcanic lakes are rare elsewhere. Birket Ram on the Golan Heights is the only crater lake in the Levantine volcanic provinces linked to Jordan rifting, while within Spain they are restricted to a cluster of relict lakes in the Campo de Calatrava of Ciudad Real, central Spain (HernándezPacheco 1932). Owing to their origins in volcanic explosions, crater lakes tend to have a small surface area and are often almost circular, with small, well-defined catchments (Timms 1992). While some are deep, oth-

262

Neil Roberts and Jane Reed

Fig. 9.2. Lake Pamvotis, Ioannina basin, north-west Greece: an example of a freshwater lake in a karstic intermontane landscape (photo: Ian Lawson). In Chapter 4, Figure 4.8 shows a SPOT image of the Ioannina basin and surrounding landscapes.

less important than in desert regions, this is a feature of some Mediterranean lakes and, in some cases, occurred at times of drier climate than at present. Wind deflation has had a significant influence on the formation of playas in the Ebro valley of north-eastern Spain and parts of central Spain (Comín and Alonso 1988; Florín et al. 1993) and the chotts and sebkhas of the Mediterranean hinterland of North Africa, for example (Chapter 14). Aeolian deflation can also remove the lacustrine sedimentary infill of dryland lake basins and produce stratigraphic records that are broken by hiatus.

Glacial Fig. 9.3. Meke Tuzlası, a hypersaline lake occupying a Late Pleistocene crater on the Anatolian plateau, with a new volcanic cone rising through the middle (photo: Neil Roberts).

ers have been infilled with sediment over time, such as the Laguna de Fuentillejo, Campo de Calatrava, or Eski Acıgöl in Cappadocia, central Turkey.

Aeolian Wind deflation can deepen and enlarge existing depressions in the landscape, or skew the shape of shallow basins in the direction of the prevailing wind. Although

The Mediterranean does not have large glacial lakes because most of the region lay to the south of the maximum extent of the major ice sheets during the last glaciation. A few lakes merit inclusion whose origins are the result of glacial scouring of pre-existing tectonic basins. They comprise the largest lakes in Italy, which are located in the Alpine foreland zone of northern Italy, and include Como, Garda, Maggiore, Iseo, and smaller lakes such as Orta. These are deep (some of their deepest parts are below modern sea level), elongate in shape and—in their northern sectors— steep-sided. In addition, the summits of many Mediterranean mountain ranges were high enough to support

Lakes and Wetlands

cirque-type glaciers during the Pleistocene (Chapter 12), and their subsequent retreat and—usually—demise, has left behind a scattering of glacial lakes. These highaltitude lakes are generally small, but can occasionally be larger, as in the Lac d’Allos in the southern French Alps, although their character is alpine rather than Mediterranean.

Coastal-fluvial The modern configuration of the Mediterranean coastline emerged only during the early to mid Holocene following glacio-eustatic rise in global sea levels (Chapter 13). Subsequently, lateral sediment transport has created beach barrier systems behind which have formed coastal lagoons, usually containing salt or brackish water. Coastal lagoons are common on lowlying coasts, especially near major river deltas such as the Nile (e.g. Bardawil lagoon), Po (e.g. Laguna Véneta), and Rhône (e.g. Étang de Vaccarès). Wetlands are, of course, an integral element of all delta and river floodplain systems, as in the Camargue (France) or the Doñana wetlands of the lower Guadalquivir (Spain), although many have been progressively drained for conversion to agricultural land. In some cases, alluvial progradation has created new lakes by cutting off former marine embayments from the sea. Bafa lake in western Turkey, for example, came into existence this way in late Roman times, as the mouth of the Büyük Menderes (Greater Meander) river prograded seawards (Müllendorf et al. 2004). Most coastal lakes are shallow, a notable exception being Skadar which straddles the border between Albania and Montenegro reaching depths of up to 350 m (Figure 9.4 and Table 9.1) One notable and surprising feature is the paucity of lakes on almost all Mediterranean islands, even large ones such as Cyprus and Sardinia. Crete supports a single small natural lake in Kournas; Sicily similarly has only one lake other than lagoons, at Pergusa (Figure 9.5), whose circular shape led to a motor race track being built around it! (Sadori and Narcisi 2001), as does the volcanic island of Pantelleria, whose lake is labelled the ‘Mirror of Venus’.

Lake Hydro-Chemistry The circum-Mediterranean lies between the humid temperate zone of overall water surplus (northern and north-western Europe) and the arid zone of severe water deficit (Sahara–Arabia). Lakes of the northern Mediterranean, those at higher altitude, or those with

263

large contributing hydrological catchments, have positive water balances and a significant outflow either as surface or groundwater discharge. Because this outflow also removes solutes, the remaining water is chemically dilute; that is, the lake is a freshwater body. By contrast, lakes in semi-arid regions, lying at low altitudes and/or with small catchments, receive less water input and lose proportionally more through evaporation from the lake surface. When there is little or no water outflow from the system and the lake is hydrologically closed, or endorheic (internally fed), then solutes are concentrated and the lake water is saline. By virtue of the linkage between a negative water balance and solute accumulation, most salt lakes tend to be shallow-water bodies. They include ephemeral hyper-saline playa-type lakes such as the Great Salt Lake (Tuz Gölü) of the central Anatolian plateau, and innumerable small, temporary salt lakes. According to figures cited in Blondel and Aronson (1999), temporary salt lakes form 11,600 km2 of the total area of 21,000 km2 covered by Mediterranean wetlands, many of which are in North Africa and Spain (Crouzet et al. 1999). Not all saline lakes are shallow however; Van in Turkey is a notable example of a large, deep salt lake. Freshwater lakes and marshes account only for approximately 2,800 km2 of the total Mediterranean wetland area, and coastal lagoons 4,700 km2 , according to Blondel and Aronson (1999). Unlike coastal lagoons, where salinity is derived from seawater and waters are dominated by Na+ and Cl− , the chemistry of athalassic (inland) salt lakes relates to the geochemistry of the catchment bedrock and to the degree of within-lake evaporative concentration (Hammer 1986). In addition, a chemically disparate variety of salts is precipitated from lake waters of varying brine composition following the model of Eugster and Hardie (1978). With increasing concentration, carbonates (least soluble) precipitate out first, often followed by sulphates and then chlorides (most highly soluble). Given the importance of calcareous bedrock, many saline lakes have evolved to precipitate marl (calcite or aragonite CaCO3 ) or—in cases where Mg is abundant—even primary dolomite (CaMg(CO3 )2 ; Irion and Müller 1968). Where volcanic or other basic rocks are present, lake water is likely to include Cl− or SO2− 4 as important anions, and in saline systems this can lead to the formation of gypsum (CaSO4 ), sodium sulphate (Na2 SO4 ), halite (NaCl), or more complex salts. Even in freshwater Mediterranean lakes, catchment geology is important in determining major ions, pH and nutrient and trace metal availability for limnic organisms.

264

Neil Roberts and Jane Reed

Fig. 9.4. Skadar, the largest freshwater lake (by surface area) in the Balkans. This coastal lake supports one of Mediterranean Europe’s most important wildfowl communities (photo: Jamie Woodward).

Spain is the only western European country to contain true inland saline lakes. The relative proportions of major ions in fifty-seven Spanish lakes spanning the gradient from fresh to hypersaline (Reed 1998) is given in Figure 9.6. Their chemistry relates closely to their origins in the dissolution of evaporites, and carbonatedominated lakes are restricted largely to the relatively few freshwater, karstic systems such as the Ruidera

(a)

system of La Mancha, central Spain. In terms of saline lakes, Figure 9.6 shows the marked contrast noted by Comín and Alonso (1988) between predominantly sodium- and chloride-dominated lakes overlying marine halite in Andalucía, southern Spain, and lakes overlying gypsum-rich continental evaporites which tend to be dominated by sulphate and calcium or magnesium in La Mancha, central Spain.

(b)

Fig. 9.5. Lago di Pergusa, a small shallow circular lake on Sicily, whose salinity and water level fluctuate seasonally in response to climate and water management. The contrasting photographs show (a) the lake full and (b) desiccating (photos: Rosa Termine (Sicilia Ambiente)).

Lakes and Wetlands

265

Na + K 100

La Mancha, central Spain Andalucía, southern Spain

0

20

80

Ebro basin, north-eastern Spain 40

60

60

40

80

20

100

0 CO3 + HCO3

100

80

60

40

20

0

SO4

Fig. 9.6. Ternary diagram showing the major chemical anion composition of fifty-seven inland lakes in Spain (modified from Reed 1998).

An excellent regional example of the relationship between hydrological balance and lake-water chemistry is provided by the Jordan River system. The Jordan forms a north–south hydrological cascade from wellwatered highlands around Mt Hermon to the arid shores of the Dead Sea at more than 400m below sea level— the lowest point anywhere on the Earth’s surface. Near the upper end of the cascade lie lakes Huleh (now mainly drained, see below) and Kinneret/Tiberias—the biblical Sea of Galilee; both lake systems are fresh with surface outflow. At the lower end of the system the Jordan River debouches into the Dead Sea (Figure 9.7), meaning that the drainage basin is endorheic. As a result the Dead Sea has accumulated solutes to the point where they are ten times more concentrated than seawater (which is typically around 35,000 ppm). With over 3 per cent of Dead Sea water in fact composed of salt, conditions are so buoyant that it is famously possible for a person to float in this inland lake without the need to swim! One of the important distinguishing features of closed, saline lakes is that they respond dynamically to changes in water balance by adjusting the size of the surface area over which water is lost by evaporation. At times of negative water balance, a closed lake will shrink in extent, its water levels will drop and its salinity will increase;

and vice versa at times of positive water balance. Water levels thus change constantly in these lakes, in response to fluctuations in climate, notably rainfall, and human impact through water use and land-cover change in the catchment. Crater lakes and other relatively ‘simple’ closed water bodies can sometimes act as giant rain gauges, whose water level and salinity variations reflect past periods of drought and flood (e.g. Jones et al. 2005). One valuable way of categorizing the water balance of Mediterranean lakes is via their stable isotope hydrology, in particular the ratio between the two principal isotopes of oxygen, 18 O and 16 O (usually represented as ‰18 O). Freshwater lakes with a short residence time have an isotopic composition similar to that of incoming precipitation, while those with a longer residence time and large evaporative losses have increasingly positive ‰18 O values lying off the regional meteoric water line (Leng et al. 1999). In the upper Jordan valley, for example, the ‰18 O of mean weighted precipitation is around −7‰, in Lake Huleh it is about −5.5‰, in Lake Kinneret it is about −3.5‰, while Dead Sea waters are around +4.5‰, reflecting progressive evaporative and isotopic enrichment downstream through this cascading hydrological system. Stable isotopes not only provide a means of classifying modern lake waters, but also

266

Neil Roberts and Jane Reed

Fig. 9.7. The Dead Sea, a large hypersaline terminal lake occupying the lowest portion of the Jordan rift. Since the 1970s, water level lowering has led to the complete separation of the shallow southern lake basin from the deep northern one (photo: Ziv Shrezer).

allow an assessment of past hydrological conditions in a region, because isotopic signatures can be well preserved as authigenic calcite-aragonite crystals or in the shells of molluscs and ostracods. In addition to salinity, the ecology of lakes is influenced greatly by the acidity (pH) of lake waters, and by their nutrient content. Mediterranean lakes tend to be markedly alkaline (pH > 8) due to the high proportion of carbonate rocks in their catchments and also highly productive. This applies both to freshwater and saline lakes, and has important implications in terms of their vulnerability to eutrophication with artificial nutrient enrichment from human activities (see below).

Lake Biology Blondel and Aronson (1999) give an excellent summary of the flora and fauna of different types of Mediterranean wetland. In contrast to oligotrophic glacial lakes, which tend to lack major growth of higher plants, many karstic or tectonic-karstic lakes, such as Prespa in Greece

and Dkada in Montenegro/Albania, are highly productive with abundant floating and emergent vegetation (Figure 9.4). They are also often important in terms of their fisheries value and as migratory and breeding grounds for birds (Chapters 5 and 23). Additionally, temporary saline lakes offer an extreme habitat to biological organisms, with major fluctuations in water availability and salinity, and periodic desiccation. As a consequence, the aquatic flora and fauna as a whole offers a biologically interesting array of organisms with special adaptations to this environment (Williams 1987). The brine shrimp, Artemia salina, for example, is the staple diet of the flamingo and can survive in salinities ranging from around 3–4 gl−1 up to hypersaline (Hammer 1986). Another striking feature of Mediterranean wetlands compared to those of temperate north-western Europe is the high level of endemism and species richness. This is in line with a general decrease in biodiversity polewards from the equator (Gaston 1996), with hotspots of plant biodiversity in Spain, Turkey, and parts of the Balkans (International Union for the Conservation of

Lakes and Wetlands

Nature, cited in World Conservation Monitoring Centre 1992) as discussed in Chapter 23. Endemism is partly a function of topographic variability and dissection by mountains, which lends itself to the isolation of populations (Blondel and Aronson 1999), and partly due to the location of the Mediterranean, and other parts of the Balkans, at a crossroads of faunal and floral influences (Kryštufek and Reed 2004; Chapter 5). Almost half the fish species of the northern Mediterranean are regional or local endemics (Crivelli and Maitland 1995). As with other ancient lakes, endemism is particularly high in Lake Ohrid (Stankovi´c 1960). Papers in Griffiths (2004) review the topic in detail for the Balkans, incorporating discussion of lakes of Greece and the southern Balkans. Endemism operates at all levels in the food chain, from fish (Banarešcu 2004) and molluscs (ibid.; Frogley and Preece 2004; Korniushin 2004), to microscopic ostracods (Griffiths and Frogley 2004) and diatoms (Reed 2004). A particularly rare biochemical environment exists at the edge of Lake Salda in south-west Turkey. Here, groundwater rich in magnesium comes to the surface and evaporates in a chemical process mediated by cyanobacteria (blue-green algae) and diatoms, to create white cauliflower-like masses of hydromagnesite up to 7m high and 200m across. These microbialites have been suggested as an analogue for conditions on Mars when primitive organisms may have flourished there billions of years ago (Russell et al. 1999). In general, the composition of the zooplankton, phytoplankton (open water fauna and flora), and benthos (shallow water bottom-dwelling communities) varies between Mediterranean lakes according to key water chemistry parameters such as total salinity, brine composition, nutrient status, and silica availability. Lake morphology is another parameter exerting a major influence on the character of the ecosystem, in determining whether or not the waters of a lake are stratified, which in turn dictates important parameters such as seasonal nutrient availability. In lakes that are deep relative to their surface area, stratification results when the uppermost water layer (epilimnion) becomes isolated from the lower water layer (hypolimnion) in times of poor mixing (low wind stress), either seasonally or year-round (Wetzel 2001). When stratified, there is a sharp gradient in water temperature (thermocline) between the two— the epilimnion being warmer and less dense–and the hypolimnion becomes starved of oxygen. These anoxic conditions effectively restrict aquatic life to the upper water body, and the absence of benthic organisms means that lake bottom sediments are undisturbed by bioturbation. (Such conditions are analogous to those in the

267

Mediterranean Sea that led to the formation of sapropels as discussed in Chapter 2.) In consequence, stratified lakes can have seasonal laminations (varves) preserved within them. A similar phenomenon can occur in saline lakes due to salinity-driven density differences, leading in this case to the stratified state of meromyxis. Several of the deep volcanic lakes of Italy and Turkey are forming varved sediments today (e.g. Albano, Nar) or have done so in the past (Mezzano, Eski Acıgöl), as are the saline lakes of Van and Burdur. The aquatic species composition of individual lakes, being closely allied to their water chemistry, can also be used as a means of wetland classification. This is most easily achieved using organisms that lie at or near the base of the trophic pyramid, notably algae (e.g. diatoms) and invertebrates (e.g. ostracods: microcrustacea). These organisms have the advantage of possessing hard parts composed of silica and calcium carbonate, respectively, that preserve after death. As a tool for quantifying past changes in water chemistry related to parameters such as climate change, scientists have developed the transfer function technique to model statistically the relationship between modern water chemistry and species assemblage composition along the environmental gradient of interest (e.g. salinity, nutrient status expressed as total phosphorus, or pH), taking modern samples from a large regional data set of lakes (Birks 1998). The technique is most advanced in diatom analysis and within the Mediterranean, salinity transfer functions have been developed for lakes in Spain (Reed 1998), Africa (Gasse et al. 1995), and Turkey (Kashima 1994). Diatom-based research into nutrient response within fresh or saline lakes is currently lacking for the Mediterranean, although data and transfer functions are available for other parts of Europe through the EU European Diatom Database Initiative (Battarbee et al. 2000). Ostracod-based salinity transfer functions have also been developed for Spain (Mezquita et al. 2005), and Turkey.

Lakes as Archives of Environmental History Long-term Climate Variability and its Consequences Because most Mediterranean lakes are of non-glacial origin, their genesis can often be traced back much earlier than those of temperate Europe and North America, in some cases (e.g. Ohrid) dramatically so. Consequently,

268

Neil Roberts and Jane Reed

some Mediterranean lake basins contain continuous or semi-continuous sedimentary records that offer very important insights into environmental and climatic conditions during Pleistocene glacial–interglacial cycles (Table 9.3; Chapter 4) as well as during the Late Glacial and Holocene (Roberts and Wright 1993; Harrison and Digerfeldt 1993). In the lowest part of the Jordan rift, lake levels rose by more than 200 m during glacial times (Marine Isotope Stages 2–4; ∼70–12 ka) to create megalake Lisan (Bartov et al. 2002). Thick laminated lacustrine Lisan beds now surround the Dead Sea and are being actively eroded by water and wind. Lakes much larger than at present also occupied other east Mediterranean endorheic basins such as Konya and Burdur in Turkey (Erol 1978; Roberts 1983). These Pleistocene mega-lakes were initially interpreted as having resulted from a climate much wetter than today’s, and led to the idea that high-latitude glacial phases were matched by low–mid-latitude ‘pluvials’. In fact it is far from certain that precipitation levels did increase in absolute terms, because the cooler climate of the Late Pleistocene also led to a reduction in evaporative water loss. This in turn would have caused lake levels to rise even without any change in the amount of rain and snow. The question of whether the enlarged glacial-age lakes of the Mediterranean resulted from a cold, dry (minevaporal) or a cool, moist climate remains a matter of debate (Prentice et al. 1992; Digerfeldt et al. 2000). Two observations may be pertinent here. First, pollen records from Mediterranean lake sediment cores show that tree cover was significantly reduced during glacial phases compared to interglacial optima (Chapter 4). This situation stands in contrast to the ‘pluvial’ lakes of the American Southwest (e.g. Lake Bonneville) where palaeobotanical evidence shows that tree cover expanded at the same time as lake levels rose (Thompson et al. 1993). Such a contrast between the two regions would appear to favour a cool, moist climatic scenario for the North American lake basins, whereas in the Mediterranean the combination of high-level lakes and Artemisiachenopod steppe seems more consistent with a generally cold but dry climate (Peyron et al. 1998; Roberts and Wright 1993; Chapter 4). Second, improved dating of Pleistocene lake beds has shown that important water-level fluctuations took place within MIS 2–4, with many basins showing maximum lake levels during the period leading up to the global Last Glacial Maximum (LGM), that is around 30–20 ka (Fontugne et al. 1999; Bartov et al. 2002; Chapter 4). Water levels subsequently fell and some lake basins dried up completely during the period 20–15 ka, when lake water balances may have been even more negative, and the

climate more intensely arid, than at present. The level of Lake Van, for example, fell by several hundred metres at this time (Wick et al. 2003). This pattern suggests that superimposed on cooling prior to, and warming after, the LGM was a climatic drying trend between ∼25 and ∼15 ka (Chapters 4 and 12). A recent synthesis of isotopic records from twenty-four lake basins around the Mediterranean has allowed a reconstruction of multi-millennial-scale trends since the global LGM. Resultant ‰18 O data on biogenic and endogenic carbonates show that isotopic enrichment was most marked during cold phases such as Heinrich Event 1 and the Younger Dryas (Roberts et al. 2008). These data indicate that Late Pleistocene cold stages in the North Atlantic region were marked by arid phases around much of the Mediterranean basin (ibid.; Chapter 4). As temperatures rose quickly around 15 ka, so too did lake water levels across the Mediterranean, reflecting an apparently synchronous increase in both temperature and humidity. Limnological conditions during the last deglacial transition (15–10 ka) are well recorded in cores from lakes in Italy (e.g. Monticchio) and the eastern Mediterranean (e.g. Zeribar; Stevens et al. 2001; Snyder et al. 2001). These show a two-step climatic amelioration interrupted by a short-lived return to lower lake levels and/or higher salinities around 13–11.5 ka. This reversal broadly coincides with the Younger Dryas climatic event, which seems to have been characterized by dry climatic conditions in the Mediterranean region. In varved sediment records, such as Van and Eski Acıgöl, the main warming–wetting transition at the start of the Holocene occurred very rapidly, in less than a century. Evidence for changes in lake-water chemistry and productivity comes from a wide variety of techniques such as diatom analysis and geochemistry, including stable isotopes. East Mediterranean lake isotope records show a consistent pattern at the Pleistocene–Holocene transition, with more positive ‰18 O values during the Younger Dryas stage, reflecting strong evaporative enrichment despite the low temperatures, followed by a shift to more negative values at the onset of the Holocene (Figure 9.8). During the Holocene, stable isotopes and other proxyclimate data from eastern Mediterranean lakes suggest more favourable water balance conditions during the early to mid-Holocene with a return to drier conditions during the late Holocene (e.g. Roberts et al. 2001; 2008). This pattern seems to be mirrored in lake records from southern Italy, such as Pergusa on Sicily (Sadori 2001) and Albano. Ariztegui et al. (2000) have correlated a period of wetter climate recorded at Lake Albano to Mediterranean deep-sea cores which show a period of

Lakes and Wetlands Eski Acigöl

Van

‰ −2 0

‰ 2

−4

−2

2

0

269

Zeribar 4

−8 −6

6

0

0

4

1

1000

5

2

2000

3

3000

4

4000

‰ −4

−2

0

6 7

7

~7000

8

5000 6000 7000 8000

10

9000 10 10000 ~12000

11000 12

~6400

11 12 13 14

9

11

Depth (m)

5 6

Varve age BP

Depth (m)

8 9

~9000

15 16 17

~13200

Fig. 9.8. Comparative oxygen-isotope curves for three east Mediterranean lake records (modified from Roberts et al. 2002; original data from Roberts et al. 2001; Stevens et al. 2001; Wick et al. 2003). The Van varve chronology underestimates the true age of the start of the Holocene by around 1 ka. In at least two of these records, the ‰18 O data reflect primarily shifts in effective precipitation from wetter (left) to drier (right).

organic sapropel deposition between 9–7 ka, caused by marine anoxia (Chapter 2). They also identify a brief interruption to this wet climatic phase linked to the 8.2 ka arid event recorded in tropical lake levels and the Greenland ice cores. While a pattern of a wetter early to mid-Holocene climate is indicated by many lakes from the eastern and central Mediterranean, those in the western part of the basin show more variable trends. Many lakes do exhibit a mid-Holocene phase of maximum lake levels, or a ‘climatic optimum’ inferred from pollen data, within the mid-Holocene (e.g. Lago de Banyoles; Pérez-Obiol and Julià 1994; Laguna de Medina; Reed et al. 2001; sites in Menorca; Yll et al. 1997), but this is not universal. There is also some evidence that the onset of the early Holocene humid phase was earlier in the east than the west (Reed et al. 2001; Valero-Garcés et al. 1999), which may reflect the differences between a predominantly Mediterranean versus Atlantic influence. However, the mechanisms are still unclear, due in part to the paucity of continuous Holocene sediment records for the western Mediterranean region. The complex palaeoenvironmental record of the shallow, saline Laguna de Medina in southern Spain illustrates these fluctuations in lake water balance and salinity during the last 9,000 years. The highest lake

levels were experienced around 7.0–6.7 ka, but the lake history was interrupted by several short-lived arid phases (Reed et al. 2001) (Figure 9.9). Abrupt arid events are also a feature of the Holocene water-level record from Lake Tigalmamine in the Middle Atlas of Morocco; interestingly, pollen data from the same cores indicate that the forest ecosystem here was able to survive through these century-scale drought episodes (Lamb and van der Kaars 1995). Similar arid events have been identified later in the Holocene elsewhere around the Mediterranean (summarized in Shennan 2003), notably at 4.2 ka (2200 BC), an event that has been linked to cultural collapse in Egypt (Hassan 1997; Stanley et al. 2003) and the northern Levant (Weiss et al. 1993). Not all ancient societies responded in the same way to drought and stress, however: the Bronze Age cultures of the southern Levant, for example, adopted irrigation and food storage and redistribution systems to cope with water shortages (Rosen 2007). Other short-lived and, in some cases, catastrophic, environmental impacts resulted from explosive volcanic events, such as the massive eruption of the Aegean island of Santorini (Thera) during the early 2nd millennium BC (Chapter 15). Distal volcanic ash (tephra) from this and other eruptions can be found in many

270

Neil Roberts and Jane Reed

a lla h te w lo octa c Cy ch

h tc

ee

an

a

u sc di lo s py eu m p Ca cly

s s ei id a pr ros Cy to

rid yp i oc ton i es w Pl ne

op

si

s

C

n he

op

i od

ac Li

ea

m

e

no

lo

c gi

al

zo

ne

d ve rre le fe e In lak

l

0

1000

5

14

C age (years BP)

2000

3000

4

4000

5000

3b 6000

3a 2

7000

1

8000 0 20 40 60

0 20 40 60 80 100

%

%

0

1000 2000 0

g dw−1

2000

g dw−1

4000 0 20 40 60

MAX.

%

Fig. 9.9. Key selected palaeolimnological indicators and inferred record of Holocene lake-level change in the Laguna de Medina, Cádiz, south-west Spain. The long-term water-level record was inferred mainly from the ecology of diatoms and ostracods. Abrupt desiccation events were indicated mainly by the pollen of Chenopodiaceae (indicating extension of shoreline vegetation), and a range of lithostratigraphic indicators (not shown) (modified from Reed et al. 2001). Nos. 1 to 5 are local pollen zones.

lake sediment sequences from the central and eastern Mediterranean, and serve as useful marker horizons. For example, over 100 tephra layers have been identified in the ∼100-ka long sediment record from Monticchio crater lake (Allen et al. 1999). The impact of ash deposition on terrestrial and aquatic ecosystems seems to have been generally slight, although fine-resolution analysis of sediments lying above the Santorini tephra layer at Gölhisar in south-west Turkey showed a probable fertilization effect on diatom algae within the lake (Eastwood et al. 2002).

Human Impact from Palaeolimnological Data From at least Bronze Age times onwards, it becomes possible to recognize unambiguously in sediment records the consequences of human action upon lake catchments, notably via deforestation that in turn led to soil inwash and changes in lake productivity and water clarity. These relationships can be studied most clearly in multi-proxy analyses of the same core archive, using pollen analysis to indicate catchment land use change, geochemistry, or mineral magnetics to highlight any

consequent erosion and sediment flux, and diatoms, ostracods, and other biotic remains to show the response of aquatic organisms (e.g. Lamb et al. 1999; Eastwood et al. 2007). In the Mezzano crater lake in central Italy, Sadori et al. (2004) were also able to compare pollen and sedimentological evidence for human impact against archaeological finds of spear heads, axes, and other metal objects, at least some of which were deliberately placed in the lake as votive offerings. Submerged lake-shore dwellings confirm the presence of human settlement dating to 3.6–3.2 ka, and also indicate that water levels were lower during Bronze Age times, in turn linked to episodes of dry climate during the 2nd millennium BC. Pollen and micro-charcoal data show partial removal of the surrounding oak-beech forest at this time, accompanied by increased landscape burning. Sediment cores from large lakes such as Van often aggregate the effects of human-induced land cover change over a large area of landscape, because their air-borne pollen catchments are similarly large. This typically results in an apparent monotonic trend of increasing anthropogenic activity towards the present day (e.g. Wick et al. 2003). By contrast, the palaeolimnological records from smaller water bodies,

Lakes and Wetlands

zo

ak

lp ca Lo

D

O

G ra

liv

ec

e

id

ss

uo

ol

us

le

n

O

ak O n ee gr er Ev

ne

Pollen

271

0

Archaeological periods

8

Israelian/Syrian/ Mandate

7

Ottoman

Magn. susceptibility composite profile

1917 AD

Mameluke/ 1516 AD Crusaders

100

6

5

Depth (cm)

200

4

1099 AD

Early Arabian

683 AD

Byzantium/ Roman 37 BC

Hellenistic 332 BC

Persian/ Babylonian

300

586 BC

Iron Age

3

1200 BC

Late Bronze Age Middle Bronze Age

400

2000/2200 BC

Early Bronze Age 3500 BC

2 500

Chalcolithic

1 0

20 40 60 0

20 40 60 80

0

20 40

0

20 40

4500 BC

0

3000

6000

Fig. 9.10. Stratigraphic changes in major pollen types from Birket Ram showing mid–late Holocene cultural and vegetation change in the Golan Heights, along with catchment erosion indicated by magnetic susceptibility (modified from Schwab et al. 2004).

such as Mezzano and Birket Ram, capture most of their pollen from the vegetation immediately surrounding the lake. In consequence, they tell a local rather than a regional story, and they show that human impact on Mediterranean landscapes during the second half of the Holocene was often episodic rather than continuous. Periods of intense human use, typically linked to the creation of agricultural landscapes, were followed by periods of partial abandonment, when there was reafforestation by secondary woodland and at least partial landscape recovery (Chapter 7). A good example is provided by the Bey¸sehir Occupation phase, originally recognized in pollen diagrams from southwest Turkey, which typically lasted from ∼800 BC to

AD ∼600 and which was followed by the development of secondary woodland vegetation (Eastwood et al. 1998). At Birket Ram in the Golan Heights, agrarian activity linked to olive cultivation can be identified in pollen samples and from archaeological evidence back to Chalcolithic times around 4500 BC (6.5 ka), but for much of the Middle–Late Bronze and Iron Ages the catchment of this Levantine crater lake seems to have reverted to deciduous oak woodland (Schwab et al. 2004) (Figure 9.10). A much larger-scale phase of forest clearance and agrarian activity began in the mid-1st millennium BC, and this was associated with accelerated soil erosion in the catchment as indicated by increased sedimentation rates

272

Neil Roberts and Jane Reed

and magnetic susceptibility values in the lacustrine sedimentary record. This interval of cultural landscape formation—broadly matching the Bey¸sehir Occupation phase—reached a peak in the Hellenistic–Roman– Byzantine era, but came to an end around the time of the Arab invasions (7th century AD), after which semi-nomadic livestock husbandry replaced olive-vinewheat cultivation as the dominant land use. Evergreen rather than deciduous oak then became the main tree pollen type, probably deriving from scrub vegetation that was able to resist grazing/browsing pressures. A similar sequence has been found in pollen cores from the Jordan valley lakes, and in those from the Dead Sea a change in sediment type from massive halite to varved aragonite/gypsum marls and back again suggests that the Roman era of agricultural extension may also have coincided with a period of more favourable climate (Heim et al. 1997). Palaeohydrological reconstructions from stable isotopes, dated lake shorelines, and marsh wetness indicators also point to the occurrence of significant climatic fluctuations during the late Holocene (Bookman et al. 2004, Vermoere et al. 2002). The ‘Little Ice Age’ (AD ∼1400–1850) for example, seems to have been drier than at present over much of the Mediterranean basin, although some mountain catchments saw glacier expansion at this time (Chapter 12) and large flood events have been recorded in river basin records (Chapter 11). In the lake sediment data from the Golan Heights, the scale of twentieth-century human impact on (semi)natural vegetation and soil erosion appears to have been, if anything, smaller than it was during Classical times. However, in other Mediterranean palaeolimnological records, it is clear that human impact on lake catchments reached a maximum during the last millennium, often during the last century. A good example is provided by the lakes of Morocco’s Middle Atlas Mountains, most of which are small and karstic. Those lakes with relatively undisturbed forested catchments, such as Tigalmamine and Azigza (Figure 9.11a), today contain clear water, with diatom algae dominated by centric plankton (mainly Cyclotella spp.), rich gastropod and ostracod faunas, and beds of charophytes and other aquatic macrophytes (Figure 9.12). By contrast, those with more than about half their watersheds deforested, such as Dayat Affourgah, Dayat er-Roumi, and Sidi Ali contain lake sediments that are today dominated by clays and other mineral matter washed in with catchment runoff (Figure 9.11b). These lakes contain mainly non-centric diatoms (typically small Fragilaria spp.) and have a much lower aquatic biodiversity (Flower et al. 1989). Cores analysed from the former group of

lakes suggest that these have maintained relatively ‘pristine’ conditions throughout the Holocene. By contrast, cores from the latter group show that most have undergone a major change in lake status, having formerly contained aquatic species similar to those found today in ‘pristine’ water bodies in the same region. The timing of this ecological transition has varied between lakes, but appears to have coincided with major land use conversion within each lake catchment as indicated by pollen analysis of the same sediment cores (Lamb et al. 1991). These combined neo- and palaeolimnological data from Morocco show particularly clearly the adverse historical consequences of large-scale logging, overgrazing, and agricultural clearance on Mediterranean wetland ecology.

Recent Trends in Mediterranean Lakes and Wetlands Resource and Conservation Status In economic terms, Mediterranean wetlands are highly valued resources for freshwater supply, notably for agriculture, recreation, and tourism, not only from lakes directly but also via exploitation of regional groundwater or surface inflows. Some lakes also have significant economic value in terms of commercial fisheries (e.g. Trasimeno, Bey¸sehir, and Dojran). Until the collapse of Yugoslavia in the early 1990s, Lake Dojran had the highest reported fisheries yield of any European lake (178.4 kg h−1 yr−1 ) (Naumovski 1991). Catches from coastal lagoons and associated wetlands are higher still, their productivity estimated as 8–10 times that of the open sea. Economic gains from lagoon fisheries along the French Mediterranean coast alone exceed those of that country’s Mediterranean trawling fleet (Blondel and Aronson 1999). The Egyptian Nile delta is another prime example; in 1995 total fish production from Egyptian freshwater fisheries amounted to 57,872 metric tonnes with a further 61,815 produced from aquaculture (Shehadeh and Feidi 1996). Fish culture in ponds accounted for two-thirds of total aquaculture production with most of the remainder coming from irrigated rice fields. Nile tilapia, carp, and mullet dominate up to water salinity values of 4 gl−1 , but up to half of existing fish farms can no longer support the first two fish due to salinization (>15 gl−1 ) of irrigation drainage water and coastal salt-water incursion, linked to the construction of the High Aswan Dam (ibid.). Another commercial aquatic resource under threat is the freshwater crayfish (Astacus leptodactylus), as a consequence of cray-

Lakes and Wetlands

(a)

273

(b)

Fig. 9.11. Lakes with contrasting catchment land uses and aquatic ecology in the Middle Atlas region of Morocco (a) Tigalmamine, a relatively pristine karstic lake set in cedar-oak forest; (b) Sidi Ali, whose catchment has been largely deforested, and whose sediments are now dominated by eroded and inwashed soils. Sidi Ali also shows a very clear former lake shoreline (bottom left) (photos: Neil Roberts).

The value of lakes as recreational or tourist attractions is often only locally significant (e.g. for camping or recreational fishing). Exceptions of national or international economic importance include the sub-alpine Italian lakes such as Como and Maggiore, and coastal lagoons offering beach tourism such as Languedoc, France,

fish plague and overexploitation. Lake E˘girdir produced over 2,000 tons of crayfish in 1984 (70% of the total Turkish catch), but this fell catastrophically to only 1 ton by 1990, with the number of active fishermen on the lake falling by 85 per cent (Magnin and Yarar 1997).

Catchment land use (modern)

Early C20th diatoms

Modern diatoms

Tigalmamine Azigza Sidi Ali Dayat Affourgah Dayet er-Roumi 0

20

40

60

80

100

Percentage

0

20

40

60

80

100 0

20

Percentage

40

60

80

100

Percentage

Woodland, marsh, and other

Planktonic Cyclotella

Arable and grazing/tree crops

Small Fragilaria

Fig. 9.12. Twentieth-century catchment land use and diatom algae species composition from the recent sediments of five Middle Atlas lakes, Morocco (data after Flower et al. 1989; Benkaddour 1993; Lamb et al. 1999). Lake sites are arranged top to bottom from least to most impacted. The early twentieth-century diatom floras from Dayat Affourgah and Dayet er-Roumi suggest that catchment disturbance has increased at these sites since that time.

274

Neil Roberts and Jane Reed

the Dalyan delta, Turkey, or the Croatian littoral. The large lakes of Dojran and Ohrid also provide important national ‘beach’ tourism facilities in the Former Yugoslav Republic of Macedonia, the only landlocked country of the Mediterranean. An emerging problem is the growth of secondary residences around lakes, particularly in Italy, or along the coastal strips of countries such as Spain and Greece. The associated environmental impact is partly visual, but also often includes water pollution from sewage and other waste emissions, and increased desiccation or salinization from accelerated water abstraction to support a growing population. In turn, these changes can lead to a decline in biodiversity (Chapter 23). Many Mediterranean wetlands are of international ecological significance, either because of the presence of endemic species, particularly in Balkan lakes such as Ohrid, or on account of their role for migratory wildfowl (Pearce and Crivelli 1994; Chapter 5). They also offer vital ecological assets by trapping or transforming sediments and nutrients and for floodwater attenuation. The ecological importance of bird populations has been a key factor in the designation of sites as nature reserves (e.g. the ducks of Laguna Salada, Cádiz, Spain (FernándezPalacios 1990), the flamingo breeding-ground of Fuente de Piedra, Málaga (García Jiménez 1993), or the wading birds of Lake Prespa (Crivelli and Catsadorakis 1997)). In Turkey there is even a special ecological designation of Important Bird Area (Magnin and Yarar 1997). Coastal and deltaic wetlands are notably important for migratory wildfowl, and include sites such as Doñana (southern Spain), the Camargue (southern France), Lake Ichkeul (Tunisia), El Kala (Algeria), the Danube delta in the western Black Sea, and lakes Burullus and Bardawil in Egypt. All of these, and more than 200 other wetland sites occupying around 7 million hectares in the circum-Mediterranean and Black Sea regions, have been designated as internationally important under the Ramsar convention (Table 9.4; Figure 9.13). This intergovernmental treaty, originally signed in Ramsar (Iran) in 1971, provides a framework for the conservation and sustainable use of wetlands and their resources (). Ramsar sites are nominated by national governments; some countries (e.g. Italy) designating a relatively large number of individually small sites, while others (e.g. Algeria) have nominated much larger areas. In contrast, other Mediterranean countries (e.g. Israel, Libya) have hardly designated any areas at all. Unsurprisingly, there are big differences between, and even within, countries in the protection given to different Ramsar sites. Thus Lake Ku¸s in Turkey’s Marmara Sea region is protected under

the national law on natural assets, but adjacent lake Uluabat is not, even though its rich aquatic flora and fauna support up to half a million waterbirds in winter (Magnin and Yarar 1997). Recent palaeolimnological work at Lake Uluabat has shown an increase in sediment input and eutrophication during the course of the twentieth century and a marked ecological change in the early 1960s leading to a sustained loss in submerged macrophytes since 2000 (Reed et al. 2008). Both sites have Ramsar designation. The conservation and biodiversity value of saline lakes in particular has often been underestimated (Williams 2000), and in some cases there are significant conflicts of interest between economic demands and the conservation of biodiversity. Such conflicts are a feature of many environments and habitats across the Mediterranean region (Chapter 23).

Current Threats to Aquatic Ecosystem Health Mediterranean lakes and wetlands are currently subject to a wide range of environmental threats, one of the most important of which is drainage or diversion for water supply and/or irrigation, including dam construction, river impoundment, and groundwater pumping (Hollis and Finlayson 1996). Over-abstraction of groundwaters has been a principal cause of the desiccation of over 50,000 ha of marshland within the Konya basin of central Turkey since 1980. An important additional element of the environmental hazards of drainage in semi-arid regions compared to temperate regions is the associated risk of salinization (Williams 2000), including seawater intrusion to coastal aquifers. The input of irrigation waters may also have extremely deleterious effects on ecosystems. These problems are often added to other more ubiquitous pollution problems affecting water quality in the developed world, such as eutrophication (nutrient enrichment) from urban and agricultural input of N- and P-compounds, and organic, pesticide, or toxic metal pollution (Mason 2002). Finally, the introduction of alien species for commercial exploitation has often been made without proper appreciation of the impact it may have on native species and biodiversity, as shown in Chapter 23. As discussed earlier, there is evidence for significant human impact on aquatic ecosystems back into prehistory from activities such as deforestation. If the consequences of human actions for lakes and other water bodies were initially unintended, deliberate transformation of wetlands had become a clear objective by historic

Lakes and Wetlands

275

Ukraine* [19] [19]

(a)

Russia* [4]] [4

Romania [2] [2] France* [3] Portugal [12]

Croatia* [2] Monaco [1]

Spain [49]

Bosnia [1]

Italy [46]

Morocco [4] Tunisia Tunisia [1]

Moldova [2] Bulgaria* [8]

Montenegro [1]

Georgia [2]

Macedonia [1] Albania [2]

Malta [2]

Israel [2]

Algeria* [19] Libya [2]

Area of RAMSAR sites (Ha)

Iran* [5]5 [

Syria [1] Cyprus [1]

Greece [10]

Armenia [2]

Tu rkey [9]

Lebanon [4] Jordan [1]

Egypt [2]

Number in brackets indicates the number of RAMSAR sites in each country * Indicates part of country only

2,500 1,000 500,0

100,0 00 50,0 10,00 00 0 <1,0 5 ,0 0 00 0

,000

,000

00

(b ) (b)

36. Trasimeno 39. Küçükçekmece 34. Tablas de Daimiel 38. Konya

33. Doñana 37. Burdur 35. Ichkeul

40. Larnaca salt salt lake lake

41. Huleh

0

1000 km

Fig. 9.13. (a) Numbers and areas of Mediterranean—Black Sea Ramsar sites by country. (b) Location of selected coastal wetlands and other threatened sites (see Table 9.4 for details).

276

Neil Roberts and Jane Reed

times. In particular the drainage of ‘pestilential’ marshes to create irrigable agricultural land was an almost obsessive Roman goal. The Pontine marshes south of Rome originally covered about 70,000 ha but the emperors Trajan and Theodoric organized the drainage of those parts lying above sea level. Work continued under plans laid down by Leonardo da Vinci in the sixteenth century AD , and the rest of the Pontine marshes was drained by a system of dykes and pumps under Mussolini’s ‘battle for the plains’. Similar large-scale ‘reclamation’ works led to the loss of wetland landscapes elsewhere in Italy, for example in the Veneto (Cosgrove 1990), Fucine, and the Tavoliere; in consequence Italy’s wetlands now cover only a tenth of their area in Roman times. Similarly, the total marshland area in Greek Macedonia was reduced from 98,600 ha to just 5,600 ha between 1920 and 1962 due to drainage (Psilovikos 1992). Lake Huleh in the Jordan rift all but disappeared between 1950 and 1958, when 5,000 ha were drained in engineering works organized by the Jewish National Fund (Hambright and Zohary 1998) (Figure 9.14). Altogether at least half of all Mediterranean marshes and other shallow wetlands have been drained in this way. Even today, wetland drainage remains an official policy in some Mediterranean countries, including Turkey, where wetland reclamation remains one of the stated functions of the state water authority (DSI). Of the 40,000 ha of wetlands lost by Spain in recent years, three-quarters were connected to aquifers. Many of the 240 small Spanish saline lakes described in a compilation by Pardo (1948) have either disappeared through drainage or been altered by the input of freshwater or urban waste (Montes and Martino 1987). Where they have not disappeared, many have switched from a perennial to a seasonal (summer-dry) status due to water level reduction (Reed 1995). Even when they are protected as important national parks, such as the Tablas de Daimiel and Doñana marshes, groundwater exploitation has continued with major associated ecological impact (Llamas 1988). There has also been a tendency for restoration programmes to focus on eutrophication rather than water shortage, in attempting to maintain habitats for breeding birds (Florín and Montes 1999). An extreme example of the conflict between water use for irrigation agriculture and wetland conservation has come in Lake Ichkeul in Tunisia, a major stopover point for hundreds of thousands of migrating ducks, geese, storks, and flamingos who come to feed and nest there. Ichkeul is one of the last remaining lakes in a chain that once extended along North Africa’s littoral and is Tunisia’s only designated Ramsar site and a National

Park (Thomas et al. 1991). During the 1980s the construction of dams on three rivers flowing into Ichkeul drastically reduced the freshwater input into this shallow lake. This development took place despite protests from national and international conservation groups, including the inscription of Ichkeul on the UNESCO List of World Heritage Sites in Danger. Unsurprisingly it has led to an increase in the salinity of the lake and marshes, and a loss of reed beds, sedges, and other freshwater plant species with associated bird species such as purple heron, purple gallinule, and reed warblers. Following a series of environmental impact assessments, further dam construction work has been halted, but this has been too late to prevent serious ecosystem damage. By a curious twist, palaeolimnological and historical studies at Ichkeul have shown that the pre-disturbance ecosystem was not in fact ‘natural’ at all, but the inadvertent result of the nineteenth-century construction of a ship canal to the coast at Bizerte (Stevenson and Battarbee 1991). The hydrological consequences of water abstraction for irrigation agriculture has been particularly dramatic for terminal (non-outlet) lakes. Thus while the diversion of freshwater from the upper part of the Jordan River system into Israel’s National Water Carrier network has not dramatically altered conditions in Lake Kinneret (an outlet lake), it has led to a sustained shrinkage of the Dead Sea at the end-point of the hydrological system. Dead Sea levels have fallen by around 0.6 m/yr−1 since 1970 (Water Data Banks Project; Bookman et al. 2004), and has led to the separation of the deep northern lake basin from the shallow southern one (Figure 9.7). The latter has now been transformed into a series of giant evaporation pans for salt extraction. A similar fate is befalling terminal Lake Burdur in south-west Turkey, where the water level fell by 11m and salinity doubled to around marine levels between 1970 and 2002, primarily as a result of dam construction and irrigation agriculture in the catchment (Figure 9.15). Burdur supports over two-thirds of the world wintering population of the White-headed duck (Oxyura leucephala), a species that is globally threatened with extinction (Magnin and Yarar 1997). But because the lake supports no commercial fisheries or significant tourism, there are no pressing economic arguments to counterbalance the overexploitation of its inflowing waters, and Burdur could eventually desiccate in the same way as the Aral Sea, with potentially catastrophic impacts for both the local human population and for the wildlife it currently supports. Another serious threat is salinization of fresh and saline lakes due to the inflow of saline irrigation water.

Lakes and Wetlands

Before 1951 Jo

rda

n

e Riv

N

r

After 1994 Jo

rda

n

Riv

277

er

Marsh Lake Agmon Einan Einan Reservoir Reservoir Lake Yesod Hama’ala

Yesod Hama’ala

3 km

River

2

River

1

Jordan

Jordan

0

Fig. 9.14. The Huleh basin, Israel, showing wetland extent before and after mid-twentieth-century drainage reclamation (modified from Hambright and Zohary 1998; see also , accessed 16 October 2008.

Effects on Nile Delta fisheries (Shehadeh and Feidi 1996) were alluded to above. Williams (2000) cites the further example of Lake Qarun, Egypt, which has switched from a fresh to a saline lake due to irrigation inflow. Apart from the obvious economic consequences in terms of freshwater supply, there are major implications in ecological terms since c.3 gl−1 salinity is an important ecological threshold (Hammer 1986). Above this level, many freshwater organisms will not survive, and the ecosystem will suffer a complete shift in species assemblage composition amongst groups of organisms, with a concomitant loss of biodiversity.

Among the more ubiquitous problems, artificial nutrient enrichment, or eutrophication, is widespread, even in large lakes often regarded as pristine, such as Lake Prespa (Greece/FYROM/Albania) (Anon. 2001), and in many saline lakes such as the Laguna de Chiprana, north-east Spain (Días 1998) or the ephemeral lakes of del Pueblo and Manjavacas in La Mancha, central Spain (García-Ferrer et al. 2003). Many of these still received direct input of sewage until very recently. Toxic metal and pesticide pollution is also common; Lake Orta, in northern Italy, for example, has suffered serious pollution by toxic metals since the installation of a rayon

278

Neil Roberts and Jane Reed

(a)

(c)

(b)

Fig. 9.15. Burdur, a saline lake occupying a tectonic basin in south-west Turkey, which has undergone a dramatic recent fall in lake level linked to over-abstraction of water for irrigation in its catchment, (a) shows the area in photograph (b); the arrows on (b) and (c) show the lake edge positions in ∼1970 and 2002, respectively; (c) shows a former boat launch station which is now some distance from the present lake edge ( photos: Neil Roberts).

manufacturing plant in 1926 (World Lakes Database); liming and other restoration attempts have not been entirely successful, and the lake also suffers contamination by PCB and organochlorine pesticides (Guzzella 1997). The banned pesticide, DDT, is still present in the sediment of Lakes Como and Maggiore, due to continued input from river inflow and sediment resuspension (Binelli and Provini 2003) and Como, despite its beauty, is locally called the ‘toilet of Italy’! Further east, the brackish-water coastal lagoon Küçük Çekmece was enveloped by Istanbul’s urban sprawl during the 1980s and 1990s, and high levels of industrial and domestic waste discharged from its catchment have made this one of the Mediterranean’s most polluted lakes (Figure 9.16). The neighbouring freshwater lagoon Büyük Çekmece seems set to experience the same fate over

the coming decades as this mega-city expands still further. While environmental degradation and overexploitation of aquatic fisheries resources may pose an economic threat, the introduction of alien species for commercial gain (or unintentionally) has often had dramatic and unpredicted ecological consequences on aquatic ecosystems in the Mediterranean. There are many examples. The North American red swamp crayfish (Procambarus clarkii) was introduced artificially for commercial exploitation in the Doñana National Park, south-west Spain in 1974, and has proved so successful that it has changed the structure of the entire ecosystem and now poses a serious challenge in management terms (Gutiérrez-Yurrita and Montes 1999). Following the introduction of pike in 1953, followed by a range of

Lakes and Wetlands

279

Fig. 9.16. Late twentieth-century urban expansion of Istanbul into the catchments of the Çekmece coastal lagoons, as mapped from ˘ remote sensing Landsat imagery (kindly provided by Dr H. Turoglu, Istanbul University). The built-up area increased more than threefold between 1975 and 1987, and by more than a factor of twelve between 1975 and 2000.

zooplanktivorous fish and, more recently, P. clarkii in the Ruidera lakes of central Spain (Elvira et al. 1996), virtually no native fish species are left in the lake and the crayfish forms the dominant prey in the restructured ecosystem. Only one of the five or six native fish species in Lago de Banyoles is still common following the introduction of twelve exotic fish species since 1910 (GarcíaBerthou and Moreno-Amich 2000), while eight native fish species have disappeared from E˘girdir, Turkey, after the introduction of pikeperch in 1955 (Magnin and Yarar 1997). In terms of biodiversity conservation, the introduction of alien species is a particularly serious concern in lakes such as Ohrid, which contain numerous endemics. In general, freshwaters are sufficiently scarce in the Mediterranean that environmental management of water resources is a key economic, political, and social issue (Chapter 21). In Spain, more than 80 per cent of drinking and irrigation water comes from governmentfunded projects (e.g. large dams and reservoirs), and

problems of water scarcity are being addressed by water transfer from areas with surplus (Estrela et al. 1996), involving major negotiation between semi-autonomous provinces. The politics extend to an international scale when lakes span national boundaries. Lake Dojran, straddling the border of the FYROM and Greece, is a prime example. Abstraction of water by Greece, which was initially granted on a fixed-volume basis by the Yugoslav authorities in 1988, has now accelerated to the point where the lake has dropped to 1.9m below the bilaterally agreed level (Griffiths et al. 2002). The lake is shallow, and the progressive reduction in water depth from 10.4m in the 1930s to around 4.8m by 1995 has destroyed lakeside tourism in the FYROM and is thought to have had a direct impact on fisheries and biodiversity. The Macedonians put the blame squarely at the feet of Greek water abstraction practices, and argue that the lake harbours a number of important endemics, including several diatom species (reviewed in Griffiths et al. 2002). In a phytoplankton study on the

280

Neil Roberts and Jane Reed

other side of the water (Temperonas et al. 2000), however, the opposite political agenda may be reflected in the authors’ contention that changes in lake-level are driven more by climate, and that the lake is dominated by cosmopolitan algal taxa which are of low conservation value. For conservation and restoration purposes, the palaeolimnological approach offers a rigorous means of determining the rate and timing of recent (post- c.AD 1850) environmental impact and, by inference, of the humaninduced causes of ecosystem degradation (Battarbee 1999). However, studies of this kind in the Mediterranean (cf. Griffiths et al. 2002) have been relatively few. Systematic monitoring of water quality has also been patchy. Lake and wetland management appears until recently to have been driven by economic incentives in its focus on water bodies which provide important sources of fresh water. A recent European Environment Agency Report on water resource problems in southern Europe (Estrela et al. 1996), for example, highlighted eutrophication problems in six lakes in Greece, but in the case of Spain and Portugal focused almost entirely on reservoirs, while no data were presented for Italy. Lacustrine (as opposed to reservoir) data for phosphorus concentrations across Western and Eastern Europe are similarly thin for Mediterranean countries in a more recent report (Crouzet et al. 1999). The approach to monitoring wetlands in European Union member countries should be improving, however, now that there is an obligation to comply with the EU Water Framework Directive; in Italy, for example, all lakes with a surface area of ≥0.5 km2 are to be monitored (Premazzi et al. 2003). This Directive applies only to the European Mediterranean and some of its impacts on river basins are discussed in Chapter 8.

Lake and Wetland Habitat Restoration Even as Mediterranean wetlands have been drained and lakes polluted, concern has been voiced by scientists and naturalists (e.g. Hollis and Finlayson 1996). In recent decades these concerns have led in some cases to a significant change in official conservation attitudes and policies. This change has been most obvious in relation to the strong control measures introduced to reduce pollution by toxic chemicals and heavy metals. In many cases, lake restoration has been effective simply by removing point sources of pollution. For example, levels of polychlorinated biphenyls (PCBs), DDT, and other pollutants in the sediments of Lake Garda, Italy, peaked in the 1970s but showed a significant reduction by the

1990s in line with usage history (Bossi et al. 1992). Similarly, controls over sewage effluent discharges have led to a reversal of the eutrophication trend (or reoligotrophication) in lakes such as Maggiore (Manca et al. 2000, Marchetto et al. 2004) or the lagoon of Albufera, Valencia, south-east Spain (Villena and Romo 2003). In other cases, however, more active interventions are required to achieve wetland restoration. In Lake Huleh (Israel), those who opposed the reclamation project had ensured that a small (350 ha) area of papyrus swampland in the south-west of the valley was set aside and this, in 1963, became Israel’s first nature reserve (Hambright and Zohary 1998). However, it did not prevent the loss of 119 animal species from the region. Subsequent shrinkage of the underlying peat created other unanticipated environmental problems, including dust storms and the release of high levels of nitrate into the Jordan River. Therefore, since 1990, a more active restoration programme has been underway, including deliberate maintenance of high water tables in the Huleh basin to ensure that the peat does not desiccate, and the creation of a new small artificial water body, Lake Agmon (Figure 9.14). These have been linked to the development of eco-tourism and a programme of environmental education. Lake rehabilitation in productive Mediterranean lakes, fresh or saline, is a complex task since they are often subject both to pollution and hydrological impacts (Reed et al. 2008). Even in terms simply of freshwater eutrophication, restoration is difficult since the ecosystem structure of productive lakes is particularly complex, and they are usually subject to a number of different point source (e.g. sewage pipe) or diffuse (e.g. agricultural runoff) inputs, and once levels of nutrients have built up there may also be significant release from the lake sediments themselves (Moss et al. 1996). The result is that removal of a major nutrient source may not solve the problem, and expensive techniques such as sewage treatment, dredging, and biomanipulation may be necessary before a lower trophic status is restored. When these problems are coupled with those of drainage or irrigation input the problems become acute, particularly in the case of fluctuating saline lakes; any restoration plan must also attempt to restore the temporal hydrological and hydrochemical dynamics (Florín and Montes 1999). The Tablas de Daimiel, central Spain, are a prime example (Bernuéz Sanz 1990b; Álvarez-Cóbelas et al. 2001). The ecologically valuable ‘natural’ hydrological state of these marshes was formerly maintained as a balance between stream and groundwater discharges, and

Lakes and Wetlands

retention by watermill dams. The marshes were subsequently reduced to one-seventh of their former extent between 1965 and 1973 through drainage, and the watermills were destroyed. The 1970s saw the start of eutrophication driven by a combination of urban and rural sources, which was then reduced in the late 1970s by installation of water treatment plants. Drainage continued, however, with the start of major abstraction for irrigation from the late 1970s to 1990s. The combined result was a 22-fold increase in Phragmites australis at the expense of rare Cladium mariscus reed beds, and significant loss of charophytes and other aquatic species. In the first major restoration project, these ecological changes were reversed to some extent by major inputs of water from a nearby aquifer, which restored aquatic plants and C. mariscus communities and reduced P. australis cover. As a separate issue, however, nutrient and zooplankton dynamics had also been affected by artificial introductions of carp, mosquitofish, and pumpkinseed sunfish, but studies by Angeler et al. (2002) involving experimental work with fish enclosures provide a potential means of predicting the outcome of future biomanipulation plans with higher confidence. The palaeolimnological record of past environmental change can also provide a powerful tool for restoration plans. In North Africa, the CASSARINA project on coastal wetlands (Flower et al. 2001 and other papers in the same journal issue) was developed to integrate extensive limnological monitoring with palaeolimnological research on the character of past hydrological and eutrophication-related impacts on biodiversity. The results highlighted the complexity of ecosystem response, but also demonstrated clearly that hydrological impacts remain the most important imminent threat. Thus, as in other parts of the world, successful restoration of Mediterranean wetlands will require an understanding of multivariate ecosystem dynamics, the development of integrated management plans that take account of water usage in the catchment as well as in the wetland proper, and an appropriate institutional framework with the necessary powers of decision. In some cases cost–benefit analysis may indeed show that rehabilitation plans are prohibitively expensive but, in general, Environment Ministries in Mediterranean countries are too often politically weak in comparison with those responsible for economic development. For example, the water level of Lake Sevan, Armenia, has been lowered by ∼20 m since 1933 as a consequence of large-scale irrigation engineering projects. There is government and international agreement about the causes and adverse environmental

281

consequences of this decline (degradation of water quality and aquatic habitat) and about what needs to be done to restore the system to hydrological balance (<www.unece.org/env/epr/studies/armenia>, 94 ff ). However, lack of finance and weak institutional capability in Armenia have meant that the implementation stages of planned restoration projects are unfinished; this includes the Vorotan tunnel to divert water into the lake, abandoned in the 1980s with only 2 km out of the total 21 km left to be excavated.

Conclusion Lakes and wetlands represent important natural capital for the circum-Mediterranean region in terms of water resources, aquatic biodiversity, and habitats for migratory birds. Many also carry major cultural and historical significance; one has only to recall the 16,000 Romans killed as they were forced to retreat into Lago di Trasimeno by Hannibal in 217 BC, or—around two centuries later—the sudden calming of the storm on the Sea of Galilee (Lake Kinneret/Tiberias) that saved Peter and his fellow fishermen from drowning. The case for wetland conservation can therefore be spiritual, symbolic, and aesthetic as well as economic and scientific. In addition, lakes preserve valuable records of climatic and environmental history, which can be compared with those from other records, such as fluvial sediments and archaeology. Together, these sources have generated valuable insights into the nature and timing of human landscape disturbance across the whole Mediterranean region. Lake sedimentary archives provide one key means of establishing long-term lake hydroecological condition and trajectory, and in identifying pre-modern base-line states, if they ever existed. In some cases these previous lake states can provide useful targets for environmental restoration, but it should be borne in mind that people have impacted upon Mediterranean lakes indirectly well back into antiquity as a consequence of catchment deforestation and land use change. If past lake states can offer a moving target in terms of future environmental rehabilitation, they are at least likely to indicate the sensitivity and robustness of each lake system to external forcing, whether natural (e.g. climate fluctuations) or anthropogenic. Some lake systems are evidently relatively robust, in a case like Lake Van because its huge water volume buffers it against disturbance. Other lakes are much more sensitive to change, but they differ in their ability to recover from perturbation. Those with rapid turnover of lake water

282

Neil Roberts and Jane Reed

are potentially self-cleansing within a few years. Lake Banyoles, for example, which was used for the rowing events in the 1992 Barcelona Olympics, has a short residence time of less than a year, with >80 per cent of its water deriving from, and being lost to, groundwater flow. In other cases, high sediment accumulation rates can lead to the burial—and hence effectively the removal—of toxins and metals within the lake-bottom muds. Perhaps the wetlands at greatest risk, other than those that have been drained completely, are first, those sensitive to impact but that also have slow rates of recovery, and second, those with high biodiversity, especially if supporting significant numbers of rare or endemic species (e.g. Ichkeul, Ohrid). Some lakes (e.g. Burdur) share both of these characteristics. The environmental future of Mediterranean lakes and wetlands will provide eloquent testament to our human ability to use these valued and increasingly scarce resources wisely and sustainably.

Acknowledgements Our thanks go to Husayin Turo˘glu, Mick Frogley, Tom Crisman, Thomas Litt, Laura Sadori, and Jamie Woodward for contributing materials directly to this chapter, and to numerous colleagues with whom we have the pleasure of visiting and researching some of the Mediterranean’s many lakes. Thanks also to Brian Rogers, Tim Absalom, and Jamie Quinn of the Plymouth Geography cartographic unit, for assistance with drafting diagrams.

References Ager, D. V. (1980), The Geology of Europe. McGraw Hill, Maidenhead. Allen, H. L. and Ocevski, B. T. (1976), Limnological studies in a large, deep, oligotrophic lake (Lake Ohrid, Yugoslavia). Archiv für Hydrobiologie 77: 1–21. Allen, J. R. M. and 14 others (1999), Rapid environmental changes in southern Europe during the last glacial period. Nature 400: 740–3. Álvarez-Cóbelas, M., Cirjano, S., and Sánchez-Carrillo, S. (2001), Hydrological and botanical man-made change in the Spanish wetland of Las Tablas de Daimiel. Biological Conservation 97: 89–98. Angeler, D. G., Álvarez- Cóbelas, M., Sánchez-Carrillo, S., and Rodrigo, M. A. (2002), Assessment of exotic fish impacts on water quality and zooplankton in a degraded semi-arid floodplain wetland. Aquatic Sciences 64: 76–86. Anon. (2001), Editorial—a call for help. The Prespa Lakes—a call for help in a distressing situation. Hydrobiologia 450: 3–4. Ariztegui, D., Asioli, A., Lowe, J. J., Trincardi, F., Vigliotti, L., Tamburini, F., Chondrogianni, C., Accorsi, C. A., Mazzanti, M. B., Mercuri, A. M., van der Kaars, S., McKenzie, J. A., and Oldfield, F. (2000), Palaeoclimate and the formation of sapropel S1: inferences from Late Quaternary lacustrine and marine

sequences in the central Mediterranean region. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 215–40. Banarešcu, P. M. (2004), Distribution pattern of the aquatic fauna of the Balkan Peninsula, in H. I. Griffiths, B. Kryštufek, and J. M. Reed (eds.), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 203–18. Bartov, Y., Stein, M., Enzel, Y., Agnon, A., and Reches, Z. (2002), Lake levels and sequence stratigraphy of Lake Lisan, the late Pleistocene precursor of the Dead Sea. Quaternary Research 57: 9–21. Battarbee, R. W. (1999), The importance of palaeolimnology to lake restoration. Hydrobiologia 396: 149–59. Juggins, S., Gasse, F., Anderson, N. J., Bennion, H., and Cameron, N. G. (2000), European Diatom Database (EDDI). An Information System for Palaeoenvironmental Reconstruction. European Climate Science Conference, Vienna City Hall, Vienna, Austria, 19–23 October 1998, 1–10. Benkaddour, A. (1993), Changements hydrologiques et climatiques dans le Moyen-Atlas marocain; chronologie, mineralogie, géochimie isotopique et élémentaire des sédiments lacustres de Tigalmamine. Ph.D. thesis, Université de Paris-Sud. Bernués Sanz, M. (1990a ), Doñana, in A. Troya Panduro and M. Bernués Sanz (eds.), Humedales Españoles en la Lista del Convenio de Ramsar. Minsterio de Agricultura, Pesca y Alimentación, Madrid, 141–50. (1990b), Tablas de Daimiel, in A. Troya Panduro and M. Bernués Sanz (eds.), Humedales Españoles en la Lista del Convenio de Ramsar. Minsterio de Agricultura, Pesca y Alimentación, Madrid, 151–8. Binelli, A. and Provini, A. (2003), DDT is still a problem in developed countries: the heavy pollution of Lake Maggiore. Chemosphere 52: 717–23. Birks, H. J. B. (1998), Numerical tools in palaeolimnology— progress, potentialities, and problems. Journal of Paleolimnology 20: 307–32. Blondel, J. and Aronson, J. (1999), Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford. Bookman, R., Enzel, Y., Agnon, A., and Stein, M. (2004), Late Holocene lake levels of the Dead Sea. Geological Society of America Bulletin 116: 555–71. Bossi, R., Larsen, B., and Premazzi, G. (1992), Polychlorinated biphenyl congeners and other chlorinated hydrocarbons in bottom sediment cores of Lake Garda (Italy). Science of the Total Environment 121: 77–93. Comín, F. A. and Alonso, M. (1988), Spanish salt lakes: their chemistry and biota. Hydrobiologia 158: 237–45. Cosgrove, D. (1990), Platonism and practicality: hydrology, engineering and landscape in sixteenth-century Venice, in D. Cosgrove and G. Petts (eds.), Water, Engineering and Landscape. Belhaven, London, 35–53. Crivelli, A. J. and Catsadorakis, G. (1997), Lake Prespa, northwestern Greece. A unique Balkan Wetland. Hydrobiologia 351: 1–196. and Maitland, P. S. (1995), Endemic freshwater fishes of the northern Mediterranean region. Biological Conservation 72. Crouzet, P., Leonard, J., Nixon, S., Rees, Y., Parr, W., Laffon, L., Bøgestrand, J., Kristensen, P., Lallana, C., Izzo, G., Bokn, T., Bak, J., and Lack, T. J. (1999), Nutrients in European Ecosystems. Environmental assessment report 4. European Environment Agency, Copenhagen.

Lakes and Wetlands Degens, E. T., Wong, H. K., Kempe, S., and Kurtmann, F. (1984), A geological study of Lake Van, eastern Turkey. Geol. Rundschau 73: 701–34. Digerfeldt, G., Olsson, S., and Sandgren, P. (2000), Reconstruction of lake-level changes in lake Xinias, central Greece, during the last 40000 years. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 65–82. Eastwood, W. J., Roberts, N., and Lamb, H. F. (1998), Palaeoecological and archaeological evidence for human occupance in southwest Turkey: the Bey¸sehir Occupation phase. Anatolian Studies 48: 69–86. Tibby, J. C., Roberts, N., Birks, H. J. B., and Lamb, H. F. (2002), The environmental impact of the Minoan eruption of Santorini (Thera): numerical analysis of palaeoecological data from southwest Turkey. The Holocene 12: 431–44. Leng, M. J., Roberts, N., and Davis, B. (2007), Holocene climate change in the eastern Mediterranean region: a comparison of stable isotope and pollen data from Lake Gölhisar, southwest Turkey. Journal of Quaternary Science 22: 327–41. Ehrlich, A. and Singer, A. (1976), Late Pleistocene diatom succession in a sediment core from Birket Ram, Golan Heights. Israel Journal of Earth Sciences 25: 138–51. Elvira, B., Nicola, G. G., and Almodóvar, A. (1996), Pike and red swamp crayfish: A new case on predator–prey relationship between aliens in central Spain. Journal of Fish Biology 48: 437–46. Erol, O. (1978), The Quaternary history of the lake basins of central and southern Anatolia, in W. C. Brice (ed.), The Environmental History of the Near and Middle East since the Last Ice Age. Academic Press, London, 111–39. Estrela, T., Marcuello, C., and Iglesias, A. (1996), Water Resources Problems in Southern Europe. An Overview Report. European Environment Agency, Copenhagen. Eugster, H. P. and Hardie, L. A. (1978), Saline lakes, in A. Lerman (ed.), The Physics and Chemistry of Lakes. Springer, The Hague, 237–93. Fernández-Palacios, J. M. (1990), Lagunas de Cádiz (Medina y Salada), in A. Troya Panduro and M. Bernués Sanz (eds.), Humedales Españoles en la Lista del Convenio de Ramsar. Minsterio de Agricultura, Pesca y Alimentación, 9–24. Florín, M. and Montes, C. (1999), Functional analysis and restoration of Mediterranean lagunas in the Mancha Húmeda Biosphere Reserve (Central Spain). Aquatic Conservation: Marine and Freshwater Ecosystems 9: 97–109. and Rueda, F. (1993), Origin, hydrologic functioning and morphometric characteristics of small, shallow, semiarid lakes (lagunas) in La Mancha, central Spain. Wetlands 13: 247–59. Flower, R. J. (2001), Change, stress, sustainability and aquatic ecosystem resilience in North African wetland lakes during the 20th century: An introduction to integrated biodiversity studies within the CASSARINA Project. Aquatic Ecology 35: 261–80. Stevenson, A. C., Dearing, J. A., Foster, I. D. L., Airey, A., Rippey, B., Wilson, J. P. F., and Appleby, P. G. (1989), Catchment disturbance inferred from palaeolimnological studies of three contrasted sub-humid environments in Morocco. Journal of Paleolimnology 1: 293–322. Fontugne, M., Kuzucuo˘glu, C., Karabıyıko˘glu, M., Hatte, C., and Pastre, J.-F. (1999), From Pleniglacial to Holocene: a C-14 chronostratigraphy of environmental changes in

283

the Konya Plain, Turkey. Quaternary Science Reviews 18: 573–91. Frogley, M. R. and Preece, R. C. (2004), A faunistic review of the modern and fossil molluscan fauna from Lake Pamvotis, Ioannina, an ancient lake in NW Greece: implications for endemism in the Balkans, in, H. I. Griffiths, B. Kryštufek, and J. M. Reed (eds.), Balkan Biodiversity: Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 243–60. Griffiths, H. I., and Heaton, T. H. E. (2001), Historical biogeography and Late Quaternary environmental change of Lake Pamvotis, Ioannina (north-western Greece): evidence from ostracods. Journal of Biogeography 28: 745–56. García-Berthou, E. and Moreno-Amich, R. (2000), Introduction of exotic fish into a Mediterranean lake over a 90-year period. Archiv für Hydrobiologie 149: 271–84. García-Ferrer, I., Camacho, A., Armengol, X., Miracle, R. M., and Vicente, E. (2003), Seasonal and spatial heterogeneity in the water chemistry of two sewage-affected saline shallow lakes from central Spain. Hydrobiologia 506: 101–10. García Jiménez, C. M. (1993), Limnología de Fuente de Piedra. Quercus, April 1993. Gasse, F., Juggins, S., and Ben Khelifa, L. (1995), Diatom-based transfer functions for inferring past hydrochemical characteristics of African lakes. Palaeogeography, Palaeoclimatology, Palaeoecology 117: 31–54. Gaston, K. J. (1996), Biodiversity. A Biology of Numbers and Difference. Blackwell, Oxford. González Martín, J. A., Ordóñez, S., and García del Cura, M. (1987), Evolución Geomorfológico de las Lagunas de Ruidera (Albacete-Ciudad Real). Estudios geológicos 43: 227–39. Griffiths, H. I. and Frogley, M. R. (2004), Fossil ostracods, faunistics and the evolution of regional biodiversity, in H. I. Griffiths, B. Kryštufek, and J. M. Reed (eds.), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 261–72. Kryštufek, B., and Reed, J. M. (2004), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht. Reed, J. M., Leng, M. J., Ryan, S., and Petkovski, S. (2002), The recent palaeoecology and conservation status of Balkan Lake Dojran. Biological Conservation 104: 35–49. Grove, A. T., and Rackham, O. (2001), The Nature of Mediterranean Europe. An Ecological History. Yale University Press, New Haven. Guilizzoni, P. and Oldfield, F. (eds.) (1996), Palaeoenvironmental analysis of Italian crater lake and Adriatic sediments (PALICLAS). Memorie dell’Instituto Italiano di Idrobiologia 55: 1–357. Gutiérrez-Yurrita, P. J. and Montes, C. (1999), Bioenergetics and phenology of reproduction of the introduced red swamp crayfish, Procambarus clarkii, in Doñana National Park, Spain, and implications for species management. Freshwater Biology 42: 561–74. Guzzella, L. (1997), PCBs and organochlorine pesticides in Lake Orta (Northern Italy) sediments. Water Air and Soil Pollution 99: 245–54. Hambright, K. D. and Zohary, T. (1998), Lakes Hula and Agmon: destruction and creation of wetland ecosystems in northern Israel. Wetlands Ecology and Management 6: 83–9. Parparov, A. R., and Berman, T. (2000), Indices of water quality for sustainable management and conservation of an arid region lake, Lake Kinneret (Sea of Galilee), Israel. Aquatic Conservation: Marine and Freshwater Ecosystems 10: 393–406.

284

Neil Roberts and Jane Reed

Hammer, U. T. (1986), Saline Lake Ecosystems of the World. Dr W. Junk, The Hague. Harrison, S. P. and Digerfeldt, G. (1993), European lakes as palaeohydrological and palaeoclimatic indicators. Quaternary Science Reviews 12: 233–48. Hassan, F. (1997), Nile floods and political disorder in early Egypt, in H. N. Dalfes, G. Kukla, and H. Weiss, (eds.), Third Millennium BC Climate Change and the Old World Collapse, Proceedings of NATO ASI Series I. vol. 49. Springer, The Hague, 1–24. Heim, C., Nowaczyk, N. R., Negendank, J. F. W., Leroy, S. A. G., and Ben-Avraham, Z. (1997), Near Eastern desertification: evidence from the Dead Sea. Naturwissenshaften 84: 398–401. Hernández-Pacheco, D. (1932), La región volcánica de Ciudad Real. Boletín de la Real Sociedad Geográfica 52: 195–213. Hollis, G. E. and Finlayson, C. M. (1996), Ecological change in the Mediterranean wetlands, in Tomàs Vives, P. (ed.), Monitoring Mediterranean Wetlands: A Methodological Guide. MedWet/Wetlands International Publications, Slimbridge UK and Instituto de Conservaçao de Natureza, Lisbon, Portugal, 5–24. Irion, G. and Müller, G. (1968), Huntite, dolomite, magnesite and polyhalite of recent age from Tuz Gölü, Turkey. Nature 220: 1309–10. Jones, M. D., Leng, M. J., Roberts, N., Türke¸s, M., and Moyeed, R. (2005), A coupled calibration and modelling approach to the understanding of dry-land lake oxygen isotope records. Journal of Paleolimnology 34: 391–411. Kashima, K. (1994), Sedimentary diatom assemblages in freshwater and saline lakes of the Anatolia Plateau, central part of Turkey: an application for reconstruction of palaeosalinity change during Late Quaternary, in D. Marino and M. Montresor (eds.), Proceedings of the Thirteenth International Diatom Symposium. Bristol, Biopress, 93–100. Kelts, K. and Shahrabi, M. (1986), Holocene sedimentology of hypersaline Lake Urmia, Northwest Iran. Palaeogeography, Palaeoclimatology, Palaeoecology 54: 105–30. Kempe, S. (1977), Hydrographie, Warvenchronologie und organische Geochemie des Van Sees, Osttürkei. Mitt. Geol. Palaeontol. Inst. Univ. Hamburg 47: 125–228. Korniushin, A. V. (2004), The bivalve mollusc fauna of ancient lakes in the context of the historical biogeography of the Balkan region, in H. I. Griffiths, B. Kryštufek, and J. M. Reed (eds.), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 219–42. Kryštufek, B. and Reed, J. M. (2004), Pattern and process in Balkan biodiversity: an overview, in H. I. Griffiths, B. Kryštufek, and J. M. Reed (eds.), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 1–8. Lamb, H. F. and van der Kaars, S. (1995), Vegetational response to Holocene climatic change: pollen and palaeolimnological data from the Middle Atlas, Morocco. The Holocene 5: 400–8. Damblon, F., and Maxted, R. W. (1991), Human impact on the vegetation of the Middle Atlas, Morocco, during the last 5000 years. Journal of Biogeography 18: 519–32. Roberts, N., Leng, M. J., Barker, P. A., Benkaddour, A., and van der Kaars, S. (1999), Lake evolution in a semiarid montane environment: response to catchment change and hydroclimatic variation. Journal of Paleolimnology 21: 325–43. Laubscher, H. and Bernoulli, D. (1977), Mediterranean and Tethys, in A. E. M. Nairn, W. H. Kanes, and F. G. Stehli (eds.), The Ocean Basins and Margins. ivA. The Eastern Mediterranean. Plenum, New York, 1–28.

Leng, M. J., Roberts, N., Reed, J. M., and Sloane, H. J. (1999), Late Quaternary palaeohydrology of the Konya Basin, Turkey based on isotope studies of modern hydrology and lacustrine carbonates. Journal of Paleolimnology 22: 187–204. Llamas, R. (1988), Conflicts between wetland conservation and groundwater exploitation: two case histories in Spain. Environ. Geol. Water Sci. 11: 241–51. Magnin, G. and Yarar, M. (1997), Important Bird Areas in Turkey. DHKD, Istanbul. Manca, M., Ramoni, C., and Comoli, P. (2000), The decline of Daphnia hyalina galeata in Lake Maggiore: a comparison of the population before and after oligotrophication. Aquatic Sciences 62: 142–53. Marchetto, A., Lami, A., Musazzi, S., Massaferro, J., Langone, L., and Guilizzoni, P. (2004), Lake Maggiore (N. Italy) trophic history: fossil diatom, plant pigments, and chironomids, and comparison with long-term limnological data. Quaternary International 113: 97–110. Mason, C. F. (2002), The Biology of Freshwater Pollution. 4th edn. Pearson Education, Harlow. Mezquita, F., Roca, J. R., Reed, J. M., and Wansard, G. (2005). Quantifying species–environment relationships in non-marine Ostracoda for ecological and palaeoecological studies: examples using Iberian data. Palaeogeography, Palaeoclimatology, Palaeoecology 225: 93–117. Montes, C. and Martino, P. (1987), Las Lagunas Salinas Españolas. Real Academia de Ciencias Exactas, Fisicas y Naturales, Madrid. Moreno-Amich, R. and García-Berthou, E. (1989), A new bathymetric map based on echo-sounding and morphometrical characterization of the Lake of Banyoles (NE-Spain). Hydrobiologia 185: 83–90. Moss, B., Madgwick, J., and Phillips, G. (1996), A Guide to the Restoration of Nutrient-Enriched Shallow Lakes. Norfolk Broads Authority. Müllendorf, M., Handl, M., Knipping, M., and Brückner, H. (2004), The evolution of Lake Bafa (western Turkey)— sedimentological, microfaunnal and palynological results, in G. Schernewski and T. Dolch (eds.), Geographie der Meere und Küsten, Coastline Reports 1: 55–66. Neev, D. and Emery, K. O. (1967), The Dead Sea. Geological Survey of Israel Bulletin 41: 1–147. Ordoñez, S., García del Cura, M. A., and Marfil, R. (1973), Sedimentación salina actual: la laguna de Pétrola (Albacete). Estudios Geológicos 29: 367–77. Pardo, L. (1948), Catálogo de los lagos de España. Biología de las aguas continentales 6. Inst. For. Inv. Exp., Spain. Pearce, F. and Crivelli, A. J. (1994), Characteristics of Mediterranean Wetlands. MedWet/Tour du Valat, France. Pérez-Obiol, R. and Julià, R. (1994), Climatic change on the Iberian Pensula recorded in a 30,000-yr pollen record from Lake Banyoles. Quaternary Research 41: 91–8. Peyron, O., Guiot, J., Cheddadi, R., Tarasov, P., Reille, R., de Beaulieu, J.-L., Bottema, S., and Andrieu, V. (1998), Climatic reconstruction in Europe for 18,000 yr BP from pollen data. Quaternary Research 49; 183–96. Premazzi, G., Dalmiglio, A., Cardoso, A. C., and Chiaudani, G. (2003), Lake management in Italy: the implications of the Water Framework Directive. Lakes and Reservoirs: Research and Management 8: 41–59. Prentice, I. C., Guiot, J., and Harrison, S. P. (1992), Mediterranean vegetation, lake levels and paleoclimate at the Last Glacial Maximum. Nature 360: 658–60.

Lakes and Wetlands Psilovikos, A. (1992), Changes in Greek wetlands during the twentieth century: The cases of the Macedonian inland waters and of the river deltas of the Aegean and Ionian coasts, in P. A. Gerakis (ed.), Conservation and Management of Greek Wetlands. Proceedings of a Greek wetlands workshop held in Thessaloniki, Greece, 17–21 April 1989. IUCN, Gland, Switzerland, 17–96. Ramsar Convention Secretariat, (2000–4), The Annotated Ramsar List. The List of Wetlands of International Importance designated by the Contracting Parties to the Convention on Wetlands (Ramsar, Iran 1971). Annotated Version. Ramsar Convention Secretariat, Gland, Switzerland. Reed, J. M. (1995), The potential of diatoms and other palaeolimnological indicators for Holocene palaeoclimate reconstruction from Spanish salt lakes, with special reference to the Laguna de Medina (Cádiz, southwest Spain). Unpublished Ph.D. thesis, University College London. (1998), A diatom-conductivity transfer function for Spanish salt lakes. Journal of Paleolimnology 19: 399–416. (2004), The potential of diatoms as biodiversity indicators in the Balkans, in H. I, Griffiths, B. Kryštufek, and J. M. Reed, (eds.), Balkan Biodiversity. Pattern and Process in the European Hotspot. Kluwer, Dordrecht, 273–90. Stevenson, A. C., and Juggins, S. (2001), A multi-proxy record of Holocene climatic change in southwestern Spain: the Laguna de Medina, Cádiz. The Holocene 11: 707–19. Leng, M. L., Ryan, S., Black, S., Altinsaçli, S., and Griffiths, H. I. (2008), Recent habitat degradation in karstic Lake Uluabat, western Turkey: a coupled limnological– palaeolimnological approach. Biological Conservation 141: 2765–83. Roberts, N. (1983), Age, palaeoenvironments, and climatic significance of Late Pleistocene Konya lake, Turkey. Quaternary Research 19: 154–71. and Wright Jr., H. E. (1993), Vegetational, lake-level and climatic history of the Near East and Southwest Asia, in H. E., Wright Jr., J. E. Kutzbach, T. Webb III, W. F. Ruddiman, F. A. Street-Perrott, and P. J. Bartlein (eds.), Global Climates since the Last Glacial Maximum. University of Minnesota Press, Minneapolis, 194–220. Black, S., Boyer, P., Eastwood, W. J., Griffiths, H., Leng, M., Parish, R., Reed, J., Twigg, D., and Yi˘gitba¸sıo˘glu, H. (1999), Chronology and Stratigraphy of Late Quaternary sediments in the Konya Basin, Turkey: results from the KOPAL project. Quaternary Science Reviews 18: 611–30. Jones, M., and the ISOMED working group (2002), Towards a regional synthesis of Mediterranean climatic change using lake stable isotope records. PAGES Newsletter 10/2: 13–15. Reed, J., Leng, M. J., Kuzucuo˘glu, C., Fontugne, M., Bertaux, J., Woldring, H., Bottema, S., Black, S., Hunt, E., and Karabıyıko˘glu, M. (2001), The tempo of Holocene climatic change in the eastern Mediterranean region: new highresolution crater-lake sediment data from central Turkey. The Holocene 11: 721–36. Jones, M. D., Benkaddour, A., Eastwood, W. J., Filippi, M. L., Frogley, M. R., Lamb, H. F., Leng, M. J., Leng, M. J., Reed, J. M., Stein, M., Stevens, L., Valero-Garcés, B., and Zanchetta, G. (2008), Stable isotope records of Late Quaternary climate and hydrology from Mediterranean lakes: the ISOMED synthesis. Quaternary Science Reviews 27: 2426–41. Roelofs, A. K., and Kilham, P. (1983), The diatom stratigraphy and palaeoecology of Lake Ohrid, Yugoslavia. Palaeogeography, Palaeoclimatology, Palaeoecology 42: 225–45.

285

Rosen, A. (2007). Civilizing Climate. Social Responses to Climate Change in the Ancient Near East. Altamira Press, Lanham, Md. Russell, M. J., Ingham, K., Zedef, V., Maktav, D., Sunar, F., Hall, A. J., and Fallick, A. E. (1999), Search for signs of ancient life on Mars: expectations from hydromagnesite microbialites, Salda Lake, Turkey. Journal of the Geological Society 156: 869–88. Sadori, L. and Narcisi, M. (2001), The Postglacial record of environmental history from Lago di Pergusa (Sicily). The Holocene 11: 655–72. Giraudi, C., Petitti, P., and Ramrath, A. (2004), Human impact at Lago di Mezzano (central Italy) during the Bronze Age: a multidisciplinary approach. Quaternary International 113: 5–17. Schwab, M. J., Neumann, F., Litt, T., Negendank, J., and Stein, M. (2004), Holocene palaeoecology of the Golan Heights (Near East): investigation of lacustrine sediments from Birket Ram crater lake. Quaternary Science Reviews 16–17: 1723–32. Schweizer, G. (1975), Untersuchungen zur physiographie von Ostanatolien und Nordwestiran. Tübinger Geographische Studien 60. Shehadeh, Z. H. and Feidi, I. (1996), Aquaculture Development and Resource Limitations in Egypt. FAO Aquaculture Newsletter14, December 1996 , accessed 16 October 2008. Shennan, S. (2003), Holocene climate and human populations: an archaeological approach, in A. Mackay, R. Battarbee, J. Birks, and F. Oldfield (eds.), Global Change in the Holocene. Arnold, London, 36–48. Snyder, J. A., Wasylik, K., Fritz, S. C., and Wright, H. E. (2001), Diatom-based conductivity reconstruction and palaeoclimatic interpretation of a 40-ka record from Lake Zeribar, Iran. The Holocene 11: 737–45. Spirkovski, Z., Avramovski, O., and Kodzoman, A. (2001), Watershed management in the Lake Ohrid region of Albania and Macedonia. Lakes and Reservoirs: Research and Management 6: 237–42. Stankovi´c, S. (1931), Sur les particularités limnologiques des lac égéens. Verhandlungen der Internationale Vereinigung für theoretische und Angewandte Limnologie 5: 158–90. (1960), The Balkan Lake Ohrid and its Living World. Dr W. Junk, The Hague. Stanley, J.-D., Krom, M. D., Cliff, R. A., and Woodward, J. C. (2003), Nile flow failure at the end of the Old Kingdom, Egypt: strontium isotopic and petrologic evidence. Geoarchaeology 18: 395–402. Stevens, L. R., Wright Jr., H. E., and Ito, E. (2001), Proposed changes in seasonality of climate during the late-glacial and Holocene at Lake Zeribar, Iran. The Holocene 11: 747–56. Stevenson, A. C. and Battarbee, R. W. (1991), Palaeoecological and documentary records of recent environmental changes in the Garaet el Ichkeul. Biological Conservation 58: 275–95. Temperonas, M., Kristiansen, J., and Moustaka-Gouni, M. (2000), Seasonal variation in phytoplankton composition and physical-chemical features of the shallow Lake Doïrani, Macedonia, Greece. Hydrobiologia 424: 109–22. Thomas, D. H. L., Ayache, F., and Hollis, G. E. (1991), Use and non-use values in the conservation of Ichkeul National Park, Tunisia. Environmental Conservation 18: 119–30. Thompson, R. S., Whitlock, C., Bartlein, P. J., Harrison, S. P., and Spaulding, W. G. (1993), Climatic changes in the western United States since 18,000 yr B.P., in H. E. Wright Jr.,

286

Neil Roberts and Jane Reed

J. E. Kutzbach, T. Webb III, W. F. Ruddiman, F. A. Street- Perrott, and P. J. Bartlein (eds.), Global Climates since the Last Glacial Maximum. Minneapolis, University of Minnesota Press, 468–513. Timms, B. V. (1992), Lake Geomorphology. Gleneagles, Glen Osmond, Australia. Valero-Garcés, B. L., Kelts, K., Delgado-Huertas, A., GonzálezSampériz, P., Navas, A., and Machín, J. (1999), Sedimentary facies analyses as paleohydrological proxies for saline lakes. Central Ebro Basin, Spain. Terra Nostra 99: 101–6. van Zeist, W. and Bottema, S. (1977), Palynological investigations in western Iran. Palaeohistoria 19: 19–85. Vermoere, M., Van Thuyne, T., Six, S., Vanhecke, L., Waelkens, M., Paulissen, E., and Smets, E. (2002), Late Holocene local vegetation dynamics in the marsh of Gravgaz (south-west Turkey). Journal of Paleolimnology 27: 429–51. Vicente, E. and Miracle, M. R. (1988), Physicochemical and microbial stratification in a meromictic karstic lake of Spain. Verh. Internat. Verein. Limnol. 23: 522–9. Villena, M. J. and Romo, S. (2003), Phytoplankton changes in a shallow Mediterranean lake (Albufera of Valencia, Spain) after sewage diversion. Hydrobiologia 506: 281–7. Water Data Banks Project, Multilateral Working Group on Water Resources, Middle East Peace Process, , accessed 16 October 2008. Weiss, H., Courty, M.-A., Wetterstrom, W., Guichard, F., Senior, L., Meadow, R., and Curnow, A. (1993), The genesis and collapse of Third Millennium North Mesopotamian civilization. Science 261: 995–1003.

Wetzel, R. G. (2001), Limnology: Lake and River Ecosystems. 3rd edn. Academic Press, London. Wick, L., Lemcke, G., and Sturm, M. (2003), Evidence of Lateglacial and Holocene climatic change and human impact in eastern Anatolia: high-resolution pollen, charcoal, isotopic and geochemical records from the laminated sediments of Lake Van, eastern Turkey. The Holocene 13: 665–75. Williams, D. D. (1987), The Ecology of Temporary Waters. Croom Helm, North Ryde, Australia. Williams, W. D. (2000), Dryland lakes. Lakes and Reservoirs: Research and Management 5: 207–12. World Conservation Monitoring Centre (1992), Global Biodiversity. Status of the Earth’s Living Resources. Chapman & Hall, London. World Lakes Database, <www.ilec.or.jp/database/eur/ eur-46. html>, accessed 16 October 2008. Yll, E. I., Peréz-Obiol, R., Pantaleón-Canó, J., and Roure, J. M. (1997), Palynological evidence for climatic change and human activity during the Holocene on Minorca (Balearic islands). Quaternary Research 48: 33–47. Zacharias, I., Bertachas, I., Skoulikidis, N., and Koussouris, T. (2002), Greek Lakes: limnological overview. Lakes and Reservoirs: Research and Management 7: 55–62. Zolitschka, B., Wulf, S., and Negendank, J. F. W. (eds.) (2000), Mediterranean lacustrine records: a contribution to the ELDP. Quaternary International 73/4: 1–144, , accessed 16 October 2008.

This chapter should be cited as follows Roberts, N. and Reed, J. M. (2009), Lakes, wetlands, and Holocene environmental change, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 255–286.

10

Karst Geomorphology and Environmental Change John Lewin and Jamie Woodward

Introduction Whilst about 12 per cent of the earth’s dry and ice-free land is covered by carbonate rocks (limestone, marble, and dolomite), the proportion is significantly higher in the landscapes that border the Mediterranean Sea. These rock types are especially widespread in the northern part of the region and limestones in particular reach great thicknesses in Spain, southern France, Italy, the Balkan Peninsula, and Turkey and in many of the Mediterranean islands (Figure 10.1). Abundant precipitation in the uplands of the Mediterranean has encouraged solutional weathering of these carbonate rocks for an extended period. The region contains some of the deepest karst aquifers in the world, with many extending deep below present sea level (e.g. Bakalowicz et al. 2008). The regional fall in base level associated with the Messinian Salinity Crisis allowed the formation of very deep, multiphase karst systems in several parts of the Mediterranean basin (e.g. Mocochain et al. 2006). Thus, karst terrains and karstic processes are very significant components of the physical geography of the Mediterranean basin. Indeed, along with the climate and the vegetation, it can be argued that limestone landscapes (including limestone bedrock coasts) are one of the defining characteristics of the Mediterranean environment. Much of the northern coastline is flanked by mountains with bare limestone hillslopes (Figure 10.2) drained by short and steep river systems whose headwaters commonly lie in well-developed karst terrain. Karst terrains are also well developed in the Levant and in the Atlas Mountains of Morocco and Algeria, while relict karst features can be identified in the low-relief desert regions of Libya and Egypt (Perritaz 2004) (Figure 10.1). Mediterranean

karst environments are also associated with distinctive soils, habitats and ecosystems as described in Chapters 5, 6, and 23. The nature and evolution of the karst landscapes across the Mediterranean region displays considerable spatial variability due to contrasts in relief, bedrock composition and structure, climatic history, and other factors. The karst geomorphological system is distinguished from other systems (e.g. glacial, fluvial, coastal, and aeolian) because of the dominant role of dissolution which results in water flowing in a subterranean circulation system rather than in surface channels (Ford 2004). In different parts of the Mediterranean basin, however, glacial, fluvial, coastal, and aeolian processes have themselves also impacted significantly upon karst landscapes. It is possible, therefore, to identify a mosaic of Mediterranean karst landscapes. This mosaic includes both active and relict terrains; with each component operating within different boundary conditions (mainly reflecting variations in climate and relief) and each with a distinctive history. In favourable settings, the precipitation of carbonate minerals from karstic waters has also produced a highly distinctive suite of features and landscapes in the Mediterranean region. Another important feature of Mediterranean karst environments is the formation of distinctive depositional environments in caves and rockshelters where important records of long-term environmental change and prehistoric human activity have been preserved (Woodward and Goldberg 2001; Bar-Matthews et al. 1999). It is also important to appreciate that limestone landscapes in the Mediterranean basin have exerted a profound influence on the imagination of humankind. For example, the legendary River Styx, issuing from the

288 John Lewin and Jamie Woodward

Carbonate rock outcrops 31 5

8

19

15

3

32 12

25

26

9 18

7

30

6

11

10 14

4

23

13

34

2

1

27

24

33 16 1. Pamukkale, Turkey 2. Arcadia, Greece 3. Trieste, Italy 4. Rock of Gibraltar 5. Grand Causses, France 6. Apulian Coast, Italy 7. Gargano Peninsula, Italy 8. Picos de Europa, Spain 9. Trebišnjica, Bosnia Herzegovina 10. Bussento, Italy 11. Dumanli, Turkey 12. Fontain de Vaucluse, France 13. Mount Tymphi and Vikos Gorge, Greece 14. Campanet Caves, Mallorca 15. Haute-Savoie, France

16. Tripoli, Libya 17. Jabal al Akhdar, Libya 18. Moracha Canyon, Montenegro 19. Dolomites, Italy 20. White Desert, Farafra Oasis, Egypt 21. Umm al Masabih Cave, Libya 22. Qatarra Depression, Egypt 23. Plitvice (River Korana), Croatia 24. Antalya, Turkey 25. Dalmatian Coast, Croatia 26. Sierra de Líbar, Spain 27. Franchthi Cave, Greece

29

17

21

28

0

500 km

28. Soreq Cave, Israel 29. Peqiin Cave, Israel 30. Argentarola Cave, Italy 31. Altimira Cave, Spain 32. Chauvet Cave, France 33. Pigeon Cave, Morocco 34. Sorbas Cave, Spain

22 20

Fig. 10.1. The distribution of the major outcrops of carbonate rocks (limestones, dolomites, and marble) in the Mediterranean region and the key sites mentioned in the text. This map does not show the full extent of karst phenomena because many limestones with well-developed cave systems are overlain by non-carbonate rocks.

Karst Geomorphology

289

Fig. 10.2. Steep limestone slopes in Kotor Bay on the coast of Montenegro. Note the steep chute and channel system and alluvial/colluvial fan on the left ( photo: Jamie Woodward).

290

John Lewin and Jamie Woodward

Gates of Hell, was reputedly based on a small river in Arcadia, Greece, with waters so cold and venomous as to prove fatal to those that drank them (Brewster 1997). The wine of Arcadia was, according to Aristotle, liable to become chalky when heated in a goat’s skin near a fire. Nearer our own time, such phenomena become comprehensible in terms of a developing understanding of snowmelt and sub-surface hydrology, stream water quality, and chemical dissolution processes affecting carbonate rocks. Ford and Williams (1989: 6) have described the Mediterranean as ‘the cradle of karstic studies’ because it was a Mediterranean limestone landscape in particular that gave enormous initial stimulus to scientific developments in geomorphology and karst studies.

The Karst of Trieste It was the karst of Trieste, a small region in the Dinaric Alps near the Adriatic coast (Figure 10.1), that played a significant role in the development of landform sciences as the field setting for the first systematic investigation of the distinctive features that have become known as ‘karst phenomena’. During his doctoral research in the karst, Jovan Cviji´c, a student of Albrect Penck in Vienna, was able to observe connections between distinctive surface forms and sub-surface structures as exposed in the recently constructed railway cuttings along the line from Trieste to central Europe. His work Das Karstphänomen (1893) is generally seen as critical in focusing and developing karst studies (see Herak and Stringfield 1972; Ford and Williams 1989). In an era strong on nomenclature and classification, a byproduct was the incorporation of local Slovene words into the scientific geomorphological literature (e.g. polje, dolina, ponor, and uvala), sometimes in German translation— indeed ‘karst’ is the German form of the Slovene ‘kras’. With numerous prefixes, the German ‘karren’ is also used for types of solution microforms on limestone surfaces, another manifestation of the strong role of German-speaking scientists in developing karst studies. Significantly, there was an early and explicit consideration of process mechanisms, and especially of hydrology and dissolution chemistry, to a far greater extent than in most contemporary landform studies in the English-speaking world which, in the late nineteenth and early twentieth centuries, were especially dominated by long-term and large-scale evolutionary models (see Beckinsale and Chorley 1991). These, too, were applied to limestone landscapes under the influence of William Morris Davis (who visited the Trieste karst

region in 1899) with the development of distinctive limestone erosion cycles.

Karst Development In general terms, karst landforms characteristically develop on mechanically strong, but chemically soluble rocks with a high degree of secondary porosity—as occurs with massive limestones (and also dolomites) that are bedded, faulted, and jointed. The uplifted carbonate rocks that dominate the main mountain chains in the Mediterranean region display many of these characteristics (Figure 10.3). Primary (intergranular) porosity in these rocks is negligible, so that water movement takes place along voids that become enlarged over time to provide conduits for water, solutes, and sediments. The hydrological cycle provides the primary source of energy for karst formation because water is the solvent that dissolves carbonate rocks and then carries the ions away in solution (Williams 2004). In broad terms, groundwater flows are driven by the difference in elevation (hydraulic gradient) between the recharge area and the discharge point or spring. Rates of bedrock dissolution are controlled by geochemical processes. In the karst system this is strongly dependent upon the extent to which the water has become acidified (by dissolved carbon dioxide) during its passage through the atmosphere, vegetation, and soil profile prior to coming into contact with the bedrock (ibid.). Each of the key karst-forming factors of hydraulic gradient, water flux, and vegetation and soil cover, vary greatly across the Mediterranean basin (as shown in Chapters 3, 6, 7, and 8). One measure of a well-developed karst environment is where the water and material fluxes in sub-surface conduits exceed those in surface channel flows. On the surface, closed depressions may develop above points of drainage input and subterranean cave development and these processes commonly include an element of mechanical failure and material collapse into voids below (Jennings 1985). Thus the characteristic landforms above and below ground that create karst terrains result primarily from the action of dissolution along the ordered pathways provided by bedrock structure (Ford and Williams 1989). Dissolution beneath a soil and vegetation cover, and on bare rock, also produces the etched, pitted, and fretted rock forms (karren) referred to above. Figure 10.4 shows a series of narrow solution channels formed in steeply sloping (almost vertical in places) limestone in the Durmitor National Park of Montenegro where rainfall exceeds

(a)

(b) Fig. 10.3. Bedding planes and joints in Mediterranean limestones. (a) Glacially-planed limestone pavement in Eocene limestones on Mount Tymphi in north-west Greece showing well-developed clints and grikes. (b) Tectonically deformed and faulted limestones with well-developed beds and joints exposed in a glaciated valley side in the mountains of Montenegro in the Durmitor National Park (photos: Jamie Woodward).

292

John Lewin and Jamie Woodward

Fig. 10.4. Large-scale solution channels in limestone in a formerly glaciated valley, Durmitor Massif, Montenegro (photo: Philip Hughes). The parallel channels have a mean width of about 12 cm and are irregularly spaced across the rock face.

3,000 mm per year. Other geomorphological processes can also be important and can operate in close association with solutional activity to breakdown bedrock surfaces. For example, Figure 10.3a shows a welldeveloped limestone pavement formed by glacial erosion where the initial joint spacings have been enlarged by solution weathering to form a regular arrangement of grikes (kluftkarren). Note also the development of smaller-scale solution flutes (rillenkarren) on the surface of the clints (flackkarren), and the important role of frost shattering in rock breakdown in this alpine environment. Carbonate rocks comprise by weight more than 50 per cent carbonate minerals and there is a continuous range between pure calcite or aragonite and dolomite (Ford and Williams 1989). It should be stressed, however, that not all limestones develop such characteristic surface and sub-surface forms. Some may be mechanically weak with high primary porosity—and this can prevent the development of major underground flow pathways. In Tunisia, for example, despite the abundance of limestones and dolomites, well-developed karst is not

widespread and Perritaz (2004) has argued that this may be due to the limited vertical thickness of the carbonate rocks in many cases. Non-limestone rocks may also develop karstic forms through dissolution: these include evaporites such as halite (rock salt) (Frumkin 1994) and sulphate rocks such as gypsum and anhydrite. These rocks are well developed in various parts of the Mediterranean region as a legacy of the Messinian Salinity Crisis (Chapter 1) and gypsum karsts in particular have been the focus of much research activity (Klimchouk et al. 1996; Calaforra and Pulido-Bosch 2003) (Figure 10.5). In addition, even some silica-cemented sandstones may produce dissolution features comparable to those observable in classical limestones. Quaternary-age cemented limestone gravels can also form striking karst landscapes, as on the Montello plateau in Italy (Ferrarese et al. 1998). Finally, it is worth noting that fluvial landforms on limestones reflect many factors common to other lithologies, as in the case of gorges transecting uplifting fault and fold mountain systems in tectonically active areas.

Karst Geomorphology

293

Fig. 10.5. Geography students from the University of Manchester exploring the cave systems in the gypsum of the Sorbas basin in southeast Spain in March 2006 (photo: Jamie Woodward). Here the main cave passages developed following a shift from phreatic to vadose conditions associated with regional base level fall in the Late Quaternary when flowing water excavated the erodible marl sediments that are intercalated with gypsum beds.

The Evolution of Mediterranean Limestones Figure 10.1 shows the distribution of carbonate rocks in the Mediterranean region. This involves, for example, a third of France and Turkey, a quarter of Spain, and 40 per cent and more of Croatia and Slovenia. With the exception of the southern and eastern Mediterranean (Libya, Egypt, and the Levant), mechanically resistant limestones commonly form high relief elements in the landscape (Figure 10.3), with eroded weaker materials removed from around them, and fault-bounded margins with deeply entrenched valleys combining to make limestone massifs a dominant landscape element in the region. At the macro-scale this can be explained in terms of geological history and tectonic evolution (Chapter 1). These limestones—including shallow lagoonal, reef, and deeper ocean facies—were formed around the open Tethys Ocean that has now been eliminated by plate movements. Interaction of the

African and Eurasian plates (involving also a series of microplates) led to the development of Alpine mountain systems in Cretaceous, Palaeogene, and Neogene times, and thus widespread folding, faulting, and uplifting of the Jurassic and later oceanic carbonate sediments (Figure 10.6). Tectonics involved early folding in the eastern Mediterranean in the Pontides and Hellenides (approximately where Turkey and Greece are now), with Palaeogene activity in the Alps and northern Spain (including the Pyrenees in particular), and Neogene activity spreading to the Betic cordillera of southern Spain and North Africa. This aspect of the evolution of the Mediterranean landscape is presented in more detail in Chapter 1. Tectonic activity continues today, especially in the north-eastern sector of the Mediterranean with subduction south of Crete and transform faulting active south of Turkey. All this led to extreme deformation in alpine areas, including the incorporation of eroded continental basement rocks in the evolving deformation zones

294

John Lewin and Jamie Woodward

Fig. 10.6. The tectonic setting for the deposition and deformation of Mediterranean limestones. (a) The present-day tectonic setting of the Mediterranean region (modified from Dewey et al. 1973; and Windley 1984; and see Chapter 1). Most of the limestone terrains shown on Figure 10.1 lie within the Alpine system. The evolution of the Tethyan Ocean and major subduction zones at (b) 200 million years BP and (c) 50 million years BP (modified from Biju-Duval et al. 1977 and King et al. 1997).

(Figure 10.6) and the formation of recumbent folds and thrusts. This activity has broadly led to the juxtaposition of rock types that can be highly contrasted physically and chemically, for example with fault-defined limestone massifs dominating the appearance of landscape and coastline on a large scale and later erosion products (e.g. flysch sediments) being removed preferentially around them (e.g. Woodward 1995; Bailey et al. 1997; King et al. 1997). Upstanding limestone blocks can form dra-

matic landscape features as in the case of the Rock of Gibraltar, with its peak at 426 m, that lies at the western end of the Mediterranean Sea (Figure 10.1). This is a block of Jurassic limestones and dolomites that has been inverted by earth movements. In the southern and eastern Mediterranean, marine onlap in Jurassic and later times (van Houten 1980) was succeeded by marine regression and the deposition of terrestrial sediments—but without the intense

Karst Geomorphology

tectonic deformation involved in Alpine mountain building that characterized much of the northern Mediterranean regions. The result has been the development of a fringe of largely undeformed limestone around the southern and eastern Mediterranean (Figure 10.1) which is different in appearance and character to those described above. Thus there are extensive and almost horizontal (low-elevation) limestone plateaus in Libya and Egypt, with deeper valleys and a more dissected landscape in the hills and mountains in the Levant. Tectonic processes have, however, been rather more active in the Levant—along the Jordan transform fault and in the mountainous Lebanon, for example, where elevations may exceed 3,000 m. By contrast, elevations on the limestone plateaus in Egypt and Libya are generally below 1000 m, with extreme aridity resulting in their partial coverage by sand seas (Waltham 2001, and see Chapter 14). These factors combine to produce a very different geomorphological setting for these karst landscapes in comparison to those in the northern and western Mediterranean. Other limestone terrains which have a somewhat lower degree of deformation and sometimes local relief also exist where they formed in continental basements or ocean floors which have been little transformed. These include parts of eastern Spain, the Grand Causses in France (with upland plateaus dissected by 500-m deep valleys), and the Apulian coast in Italy. Parts of the Apulia platform have preserved a relatively undisturbed (>5,000 m) sequence of Mesozoic limestones as the host microplate became sandwiched between the closing African and European continents (Ager 1980). The Gargano Peninsular (Figure 10.1) has the greatest development of karst features in Italy (Herak and Stringfield 1972: 103). Whilst most Mediterranean limestones are Mesozoic and Cenozoic (Jurassic and later) in age, there are some areas of Palaeozoic limestones that can give rise to notable karst landscapes. For example, Carboniferous limestones in the Picos de Europa form high mountains (rising to 2,426 m) with karst features rivaling any in Europe (Smart 1986; Gale and Hoare 1997) (Figure 10.1 and Chapter 12).

Mediterranean Karst Systems General features involved in evolving karst systems are encapsulated in Figures 10.7 and 10.8. These are based on Ford and Williams (1989) and Gillieson (1996) and the reader is referred to these and other general or specific texts for a comprehensive coverage of karst

295

geomorphology, hydrology, and geochemistry (see also Jakucs 1977; Jennings 1985; White 1988; Gunn 2004). Such features characteristically develop on limestone terrains in humid or seasonally humid environments, with sub-surface drainage developing between points of influent and effluent flow. This may involve allogenic (derived from outside the limestone area) rivers in blind valleys (with no surface topographic exit) entering a cave system that develops over time as passages are enlarged and adjacent rocks are lowered. Cave systems may develop by gravity water flow in an aerated zone (vadose system) or below a water table (phreatic system) which may involve deep water circulation under pressure (Figure 10.7). The cave systems below the surface reflect this in their morphology, with fissure-related open chambers and canyonlike forms with incising stream beds in the vadose zone, and solution tubes enlarged around their whole perimeter by upward as well as downward flow below. Figure 10.7 illustrates the evolution of a system primarily by phreatic streams, initially involving deep, water-filled (bathyphreatic) tubes, but with the later active water system becoming focused on developments at the water table, leaving the relict passages developed earlier as part of an explorable system. Water re-emerges—via the resurgence—as a major river. Modern research suggests that cave enlargement affects only some pre-solutional openings, especially drawing from diverted influent waters from perched streams, glacial meltwaters, and soluble/insoluble rock contacts (Palmer 1991).

Mediterranean Karst Springs Karst groundwaters emerge at the surface as springs. In contrast to springs in porous, non-karstic rocks, karst springs tend to have much higher discharges because sub-surface flow velocities can be high and they represent the output point from an extensive network of groundwater conduits (Smart and Worthington 2004). Table 10.1 lists twenty-nine of the largest karst springs so far identified around the world and more than half of these are from the Mediterranean region. The three largest springs in the region are from Trebišnjica in Herzegovina, Bussento in Italy, and Dumanli in Turkey with mean annual discharges of 80, >76, and 50 m3 s−1 respectively. It is important to appreciate that flows from springs can be perennial, seasonal, or intermittent and most Mediterranean systems follow the seasonal rhythm of the rainfall regime. In all aquifer systems, the larger the body of water relative to aquifer flow, the greater the residence time, and the more consistent the spring

296

John Lewin and Jamie Woodward

Fig. 10.7. The development of vadose and phreatic cave systems in a karst drainage system. The dashed line marks the level of the water table over time. In general, the main conduits or pipes become enlarged over time and the submerged (phreatic) system in (a) becomes a vadose system in later stages (b to d) as the water table falls in accord with external base level change. The karst drainage network develops vertically over time as base level falls. The Mediterranean has some of the deepest karst drainage systems in the world with very long vertical shafts. Some of the deep karsts developed during the Messinian Salinity Crisis (Chapter 1). Note how the number of penetrated fissures in the karst rock increases over time (modified from Ford and Williams 1989). See text for further explanation and compare to Figure 10.8.

discharge (ibid.). Springs that rise from depth because of the internal hydraulic head in the aquifer, are termed Vauclusian springs after the Fontaine de Vaucluse in southern France which has a mean discharge of c.21 m3 s−1 (ibid.). In upland Mediterranean karsts, the recharge area can be snow-covered throughout the winter and spring, resulting in a significant lag before the snowmelt component enters the karst aquifer. In the deep limestone karst system of Mount Tymphi in north-west Greece, for example, the amount of snowfall is an important control on the magnitude of summer flows in the Voidomatis River which is fed by a large exurgence in the Vikos Gorge almost 2,000 m below the highest peaks (Bailey et al. 1997) (Chapter 11).

The widespread occurrence and thickness of carbonate rocks in the Mediterranean region means that the karst system is a key influence upon river flows and water quality. High-intensity precipitation and snowmelt events can overload the karst system and produce surface flooding. Bonacci et al. (2006) describe a flash flood event in the Dinaric karst of Marina Bay on the central coast of Croatia. During extreme events, groundwater levels in parts of the Dinaric karst can rise at a rate of >30 m per hour and this can lead to rapid groundwater breakthrough, often under pressure, in many unexpected locations. On 6 December 2004 a precipitation event, with an estimated return period of between 800 and 1,000 years, fell on the area producing a

Atmospheric inputs Water, carbon dioxide, solutes, magnesium, calcium, potassium, sodium, chloride, sulphate, dust, smoke, lead-210, caesium-137

Allogenic inputs Water, humic acids, carbon dioxide, pollutants, gravels, suspended sediments, humus, uranium salts

Epikarst Solution doline

Sink

Percolation water

Karren Collapse doline

Water Solutes Humic acids Fulvic acids

Tree roots Relict clay fills

Uranium salts

Rockshelter

Bats Guano

Gravel Cave breakdown

Calcite speleothems

Autogenic inputs Cave breakdown, authigenic fluvial sediment, calcite, gypsum, phosphate minerals

Spring Travertines tufa Pond deposits (fine clays) Hydrothermal input

Groundwater

Cave system outputs Water, solutes, esp. calcium, magnesium, bicarbonate, fine sediments, organic acids

297

Fig. 10.8. A typical karst system in the Mediterranean region showing material inputs, stores, and outputs and associated processes in the vadose and phreatic zones (modified from Ford and Williams 1989 and Gillieson 1996). The phreatic zone lies below the water table (dashed line). Active tectonics in the Mediterranean can produce hydrothermal inputs to karst systems. Compare to Table 10.3. Note the rockshelter and cave entrance environments on the right of this karst system. A more detailed illustration of processes in these environments is given in Figure 10.18b.

Karst Geomorphology

Spar calcite

Fluvial sediments

Entrance facies Cave breakdown, fluvial and hillslope sediments, calcite, bone, pollen, phosphate minerals

298

John Lewin and Jamie Woodward TABLE 10.1. Large discharge springs of the world with flows >20 m3 s−1

Spring Tobio, Papua New Guinea Matali, Papua New Guinea Trebišnjica, Herzegovina Bussento, Italy Dumanli, Turkey Galowe, Papua New Guinea Ras el ain, Syria Tisu, China Ljubljanica, Slovenia Stella, Italy Ombla, Croatia Chingshui, China Spring Creek, Florida Oluk Köprü, Turkey Timavo, Italy Frio, Mexico Yedi Miyarlar, Turkey Mchishta, Georgia Buna, Herzegovina Coy, Mexico Liu longdong, China Kirkgozler, Turkey Silver, Florida Rainbow, Florida Vaucluse, France Sinjac (Piva), Montenegro Bunica, Herzegovina Grab-Ruda, Croatia Trollosen, Spitzbergen

Discharge (m3 s−1 ) >100 90 80 >76 50 40 39 38 39 34 34 33 33 >30 30 28 >25 25 24 24 24 24 23 22 21 21 20 20 20

Catchment area (km2 ) — 350 1,140 — 2,800 — — 1,000 1,100 — 600 1,000 >1,500 >1,000 >1,000 >1,000 >1,000 — 110 >1,000 — — 1,900 >1,500 1,100 500 510 390 —

Note: More than half of these are from karst systems in the Mediterranean region. Source: Modified from Ford and Williams (1989) and Milanovi´c (2000).

large number of new high discharge springs. Many house cellars were flooded by the rapidly rising groundwater and the main street in Marina was flooded by an open watercourse to a depth of 0.5 m. Such an event had never been recorded previously and it caused damage estimated at US$1.5 million (ibid.). Jourde et al. (2007) have demonstrated the importance of contributions from the karst groundwater system to flood flows in the Coulazon River near Montpellier in southern France.

Mediterranean Cave Systems Figure 10.8 shows a mature cave system in a karst landscape and the text boxes show the inputs, storages, and outputs of materials and some of the processes involved. Reworked soils such as terra rossa are a common input to Mediterranean karst cave systems and the root systems

of olive trees commonly exploit the cavities in the upper part of the bedrock surface. A range of surface limestone features (epikarst, dolines, and karren), and the transformations that are likely to occur as cave systems become abandoned by flowing water are also shown in Figure 10.8. These include speleothem formation (Figure 10.9a) and the partial blocking of conduits by fine clastic sediments. These sediments are often derived from non-limestone source rocks such as flysch and sandstones and they can exit the karst system via flowing water (Figure 10.10). Chemical and physical sedimentation of this kind within the karst system can permit the development of an environmental record (including the preservation of a human occupation history in cave mouth and rockshelter settings) which is not available in other eroding landscapes. The environmental records from speleothems are discussed in more detail below. Large passages in mature karst systems can transmit large flows during floods and the flux and storage of well-rounded gravels can be an important part of the clastic sediment budget. Angular, coarse-grained sediments may be introduced via dolines and roof collapse and they can accumulate in rockshelter and cave-mouth environments where physical weathering of the bedrock walls and ceiling take place. Outside the subterranean karst environment, precipitated and cemented carbonates including travertines and tufas (Figure 10.9b) may preserve a valuable and datable environmental record that is also not available for other lithological environments (Pena et al. 2000; Glover and Robertson 2003). More recent research has also revealed the abundance of limestone cave and cavern systems in the Mediterranean region, with exploration demonstrating their extreme depth dimension in many cases. Herak and Stringfield (1972) usefully catalogue the cave systems of France, Italy, and the former Yugoslavia, though the picture changes as exploration unfolds. Klimchouk (2004) has recently compiled a list of the deepest caves in the world and, as he points out, many are located in the young mountain belt that extends across Europe from the Pyrenees to the Caucasus. Many are located within the thick carbonate rocks of the countries that border the Mediterranean Sea including the Réseau JeanBernard (depth 1,602 m) in Haute-Savoie, France, and the Torca del Cerro (depth 1,589 m) in the Central Massif of the Picos de Europa of Spain. Table 10.2 lists some world comparisons to provide global context for the Mediterranean examples. The Picos contains the largest-known concentration of deep caves in the world (Rossi 2004) with at least fifteen caves more than 900 m

Karst Geomorphology

(a)

299

(b)

Fig. 10.9. Features produced by the precipitation of calcium carbonate in karst environments in Majorca, Spain. (a) Speleothems from Campanet Cave and (b) tufa deposition on a fountain in the grounds of the monastery at Luc (photos: Jamie Woodward).

deep. Many consist of fault-guided shafts interconnected by narrow canyons. The Picos were glaciated on numerous occasions during the Pleistocene and there is evidence of a large (c.50 km2 ) ice cap on the western Massif. Here many of the large vadose caves may be partially sub-glacial in origin (Rossi 2004) and Smart (1986) has pointed out that many large shafts probably owe their formation to large influxes of meltwater since they are located in areas that receive little modern drainage. A key feature of Mediterranean karsts is their extensive vertical development and this is due to the region’s distinctive geological and geomorphological history. Many karst aquifers across the northern part of

the region therefore contain multiple levels of passages and these are often joined by very deep vertical shafts (e.g. Waltham 1978; Courbon and Chabert 1986). In fact, Mediterranean karsts contain some of the deepest vertical shafts so far identified and, as far as water flows are concerned, this can result in sub-surface drainage systems with highly unusual three-dimensional configurations that are uncommon or impossible in surface catchments (Smart and Worthington 2004). Bakalowicz et al. (2008) have conducted hydrogeological assessments of deep karst systems in Lebanon. They represent very significant water resources—the Zarka system, for example, which is the source of the Orontes River, has a storage capacity of 27 billion m3 .

300

John Lewin and Jamie Woodward

Fig. 10.10. Fine-grained sediment outputs from a Mediterranean karst system. The large conduits are formed in Late Eocene limestones in the Lower Vikos Gorge of the Voidomatis River basin. The fine sediments are derived from flysch rocks and soils that overlie the limestones and they represent an important part of the clastic sediment budget in the subterranean karst system (see Figure 10.8 and Woodward 1997) (photo: Jamie Woodward).

Mediterranean Karst Landscapes Figure 10.11, based on Jakuc (1977), suggests an altitudinal range in processes and four main landscape elements (A to D); from alpine karst (with glaciers and bare limestone karst present above around 3,000 m)

through zones of increasing karstification and the development of surface forms and weathering residues at lower elevations. Glaciers are present in the region today in the highest mountains and were much more extensive during cold stages of the Pleistocene (Chapter 12). Jakuc’s model envisages a deep (vadose) cave system

Karst Geomorphology TABLE 10.2. The twenty deepest caves in the world Rank 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Name

Country

Depth (m)

Gouffre Mirolda Krubera (Voronja) Lamprechstofen Réseau Jean-Bernard Torca del Cerro Cehi 2 “la Vendetta” Sarma Vjacheslava Pantjukhina Sistema Cheve Sistema Huautla Sistema del Trave Boj Bulok Puertas de Illamina Lukina Jama Evren Gunay düdeni Sneznaja-Mezhennogo Réseau de la Pierre Saint-Martin Siebenhengste Slovacka jama Cosa Nostra Loch

France Georgia Austria France Spain Slovenia Georgia Georgia Mexico Mexico Spain Uzbekistan Spain Croatia Turkey Georgia France-Spain Switzerland Croatia Austria

1733 1710 1632 1602 1589 1533 1530 1508 1484 1475 1441 1415 1408 1392 1377 1370 1342 1340 1320 1291

Note: Most of the caves in this list were explored ‘top-down’ from an upper entrance or entrances and their depth is defined as the vertical range between the altitude of the highest entrance and lowermost point reached in the cave (see Klimchouk 2004). Some caves extend below the local water table and are waterfilled. This makes exploration difficult and some of the depths are therefore minimum values. Source: Modified from Klimchouk (2004).

in high relief zones of intermediate elevation, with an enlarged water table cave system below. The model is a useful one that encompasses many of the karst environments encountered in the Mediterranean region. It does not, however, include desert karst environments and this distinctive setting is described more fully in a later section. The impact of humans on karst environments also varies to some extent with elevation. In many parts of the Mediterranean karst surfaces have been directly modified to enhance their agricultural potential (Gams et al. 1993). Figure 10.12 shows two forms of karst surface modification that are widespread in the Dinaric karst. The left hand diagram (1a and 1b) shows how stone block-fronted terraces can be developed by removing limestone blocks to create a stepped hillslope profile and a deeper soil mantle. The right hand diagram (2a and 2b) shows how a funnel-shaped doline has been modified (again with the movement of both bedrock blocks and soil) to create a larger and flatter area for cultivation. These direct modifications to hillslopes and soils are most common in karst zone D of the Jakuc model (Figure 10.11) and further examples are presented in Gams et al. (1993).

301

The range in altitude and relief across the Mediterranean is believed to have a very strong effect on karstic processes and landscape development. Parts of Galicia (north-west Spain), the Pyrenees, the Alps, and the southern Adriatic coast receive over 1,500 mm in the mountains in the form of snowfall and annual precipitation can exceed 4,500 mm in the mountains of Montenegro. In the mountains, the hydraulic gradients in the karst system are much greater and the fall in elevation between groundwater recharge areas and outlet springs can be >2,000 m. High precipitation in these areas results in a much greater overall throughput of water and solutes so that karstic processes are much more active than in the low relief dryland karst of the southern Mediterranean where hydraulic gradients are much shallower and water fluxes are much lower. In Libya and Egypt, only very limited parts of the coastal limestone zone around Tripoli and the Jabal al Akhdar uplands receive more than 250 mm in winter rains (Chapters 3 and 11). While carbonate solubility is, in fact, greater at low temperatures, this effect is almost totally outweighed by the influence of organic acids in soil and vegetation covered areas. For the reasons discussed above, Mediterranean karst is more varied and more complex to interpret than early work in the karst plateau of Trieste was able to reveal. Even features once taken as ‘textbook’ karst forms turn out to be more varied. For example, poljes—which are flat-floored depressions up to 400 km2 and more in area—were once regarded as an end stage enlargement of cyclical surface karst development of doline-style phenomena. They are found widely in the Dinaric limestones (Figure 10.13a), but they have a variety of origins involving structural and process causes (Figure 10.13b). Thus, of the forty-two poljes identified by Gams (1978) in the former Yugoslavia, most are of border (eighteen) or structural (nine) type, rather than base level ones developed in simple limestone (one) (Figure 10.13). Many are, in fact, composite features (eighteen) as Ford and Williams (1989) have pointed out.

Mountain Karst Environments Many upland limestone areas in the Mediterranean contain deep valleys formed by glacial action and spectacular gorges produced by fluvial incision as in the Italian Alps, the Pyrenees, the Pindus Mountains, and the Taurus Mountains of Turkey (Chapter 1). Figure 10.14 shows two deep gorges in limestone on the Balkan Peninsula. In some reaches, the Moracha River gorge is extremely narrow—with no long-term fluvial or

302

John Lewin and Jamie Woodward

A

B

3000m

C D

2000m

1000m

Sea level

C – hill karsts above 1000 metres bearing plant and soil cover

B – subnival highmountain barren karsts above 2000 metres

D – hill karsts between 0 and 1000 metres

Intensity

A – mountain surfaces above 3000 metres

1

2

3

4

5

6

Legend: 1 2 3 4 5 6

1 – – – – – –

2

3

4

5

6

1

2

3

4

5

6

1

2

3

4

5

6

intensity of karstification depth of corrosion intensity of calcareous tufa deposition and dripstone formation in caves formation of furrow lapies formation of root lapies intensity of doline formation

Fig. 10.11. The vertical zonation of karst landscapes and processes in ‘European folded mountain regions’ modified from Jakucs (1977). The histograms show the intensity of a range of karst processes in the four main landscape elements.

colluvial sediment storage. The steep active channel zone occupies the valley floor and the dominant longterm process here appears to be vertical incision driven by uplift with the hard limestone bedrock resisting lateral erosion of the gorge walls. In contrast, the Vikos

Gorge example represents a more advanced stage of gorge development. This gorge system is much wider and colluvial processes have attacked the gorge walls to produce the thick talus deposits. The valley floor is also much wider with storage space for fluvial sediments of

Karst Geomorphology

1a

303

2a

2b 1b

Fig. 10.12. Two forms of intentional human modification to hillslopes in karst environments in the Dinaric region to create deeper soils and a flat surface suitable for cultivation (after Gams et al. 1993). 1a and 1b show the creation of flat agricultural land with stone-walled terrace fronts built from the removal of large blocks of limestone from the original hillslope. 2a and 2b show the creation of a surface suitable for cultivation following the infilling of a funnel-shaped doline with soil and the removal of small-scale limestone pinnacles.

Pleistocene and Holocene age (Woodward et al. 2008). The coarse-grained screes also transmit groundwater flows and cementation by karstic processes is widespread (Bailey et al. 1997). Many surface karstic features are most clearly seen in high plateau areas where they can be usefully mapped using remotely sensed imagery (Woodward et al. 2004). Alpine karst, in places involving interaction between direct Pleistocene glaciation and karstification (with tillfilled dolines and glacially scoured bedrock pavements, for example) has been explored in the Picos de Europa, Spain (Smart 1986), the Dolomites of Italy (Ferrarese et al. 1998), the High Atlas of Morocco (Perritaz and Monbaron 1998) and the Pindus Mountains of northwest Greece (Woodward et al. 2004) (Figure 10.3a). These areas are all over 2,000 m in elevation (and snow covered for many months of the year) and are subjected to physical and chemical weathering of limestones with karst features developed upon thick limestone sequences. This bare ‘high and cold karst’ (with surface valleys, dolines, and smaller-scale karren features) dominates the headwater basins of many Mediterranean river systems and provides a marked contrast to the regolith-covered karst of lower altitudes (Figure 10.11). The glaciated mountains of the Mediterranean contain extensive and very wellpreserved limestone pavements (Figure 10.3a). Some of these have been shown to be Middle Pleistocene in age and they have been subjected to several cycles of cold climate thermal shattering and active karst development during warmer climates (Hughes et al. 2006).

Recent work in the Pindus Mountains of north-west Greece has used uranium-series dating to establish the age of secondary calcites that have formed within the coarse-grained glacial sediment (till) matrix (Woodward et al. 2004). Detailed geomorphological mapping and formal lithostratigraphic assessment of the glacial sediments and landforms in the Pindus Mountains has led to the recent development of a new chronostratigraphic model for Pleistocene glacial activity in the Mediterranean region (Woodward et al. 2004, Hughes et al. 2006). Modern glacier dynamics and the Pleistocene glacial record in the limestone uplands of the Mediterranean are discussed more fully in Chapter 12.

The Desert Karst Jennings (1985) has argued that hot desert environments are not conducive to karst development because of the low rainfall inputs, very limited vegetation cover and therefore low soil carbon dioxide levels and high rates of evaporation. The presently ‘arid karst’ of North Africa is beginning to be explored (Albritton et al. 1990; Smith et al. 2000; Waltham 2001). Here the interaction is between limited present-day active development of arid micro-karstic features, active warm desert processes (especially thermal shattering, aeolian activity, and etching and scouring of limestone surfaces by sand blasting and burial of karst forms under mobile sand sheets) (Chapter 14), and what appear to be much more intensive karstic processes under former (wetter) climatic conditions (such as the early Holocene pluvial) and lower sea levels. Thus, sinkholes and caves are present along

(a)

Small and large poljes

(b)

Karstic limestone

(i) Border polje

Other karst forms, mainly loess

(ii) Structural polje

(iii) Base level polje High/low water table

0

250 km

Permanent flow

Intermittent flow

Fig. 10.13. (a) The distribution of poljes in the Dinaric karst region. (b) The formation of three types of poljes under varying structural and hydrogeological conditions (modified from Gams 1978).

Karst Geomorphology

the limestone plateaus of the Egyptian and Libyan desert (Gale et al. 1993), with curious features such as white chalk pinnacles (mostly 2–4 m high) emergent from a cover of orange sand, which are believed to be karst products (Waltham 2001) (Figure 10.15). Thousands of these pinnacles are present in the White Desert of the Farafra Oasis and small areas of karst towers (10–15 m high) are also found. Their origin is uncertain, but Perritaz (2004) has argued that these are relict karst features that formed under wetter climates in the Late Tertiary. Gypsum karst is well developed in Libya and extensive cave systems have been mapped. The Umm al Masabih Cave is 3,593 m long and it conveys ephemeral stream flows and sediments following desert flash floods (Perritaz 2004). Limited understanding of these desert karst systems means that it is often difficult to recognize the extent to which karstic features are inherited from previous climatic regimes (relict forms or palaeokarst) during wetter phases of the Quaternary or during the Late Tertiary. The extensive Qatarra depression (about 300 × 145 km) (Figure 10.1) may be a former river valley, subsequently dismembered by Late Miocene (Messinian) karstic processes and deepened by deflation. It is difficult to establish reliable geochronologies for many relict karst landscapes in the present arid zone because the secondary carbonates commonly lie beyond the range of uranium series dating. Karstic features are also developed at the coast of North Africa down to130–150 m below sea level. The Jabal al Akhdar in eastern Libya (Figure 10.1) is a large karst area with large sinkholes and cave systems that discharge into the Mediterranean Sea. Many of these features were first recorded during the Roman Period. Clearly there is scope for research on possibly distinctive systems, perhaps with some analogy to be made to the karst of the Nullabor Plain in Australia, a ‘dryland’ karst with dolines and a modest density of caves, but an absence of sinks and poljes (Jennings 1985). Speleothems are also poorly developed in this environment (Lowry and Jennings 1974). Similarly, we need more information on the processes and products of chemical weathering of karst surfaces in dryland environments. Under strong evaporation the formation of crusts and precipitates (e.g. calcretes) may have a significant impact on the development of karstic features.

Tufas and Travertines An area of strong recent interest concerns tufa and travertine deposits, which are carbonates precipitated

305

from solute-charged karstic waters (Pentecost 1995; Ford and Pedley 1996; Glover and Robertson 2003). Many authors use the term ‘tufa’ for cool-water carbonates and travertine for hot (>20◦ C) spring deposits (see Glover and Robertson 2003). Both types are common in the Mediterranean region reflecting the widespread presence of carbonate host rocks with welldeveloped karstic drainage—and the particular circumstances where active tectonic settings allow such groundwaters to be heated by geothermal energy. The Mediterranean region contains some of the best-known examples of Pliocene and Quaternary age and modern tufa deposits because the physical geography of many localities in the region yields optimum conditions for tufa formation (Figure 10.16). These processes take place at a range of scales. Large tufas may form in tectonically controlled depositional settings that allow extended periods of tufa deposition (Glover and Robertson 2003). These may incorporate many facies types with rich fossil assemblages and, because they can be dated by uranium-series methods, they may represent very significant archives of environmental change. The coastal city of Antalya and its hinterland in south-east Turkey (Figure 10.1) sits on one of the largest (c.600 km2 ) and thickest (up to 250 m) tufa deposits in the world (Glover and Robertson 2003) (Figure 10.16a). These workers argue that this large tufa complex was deposited in a large tectonically controlled basin between c.2 and 1.5 million years BP within a series of small lakes when the karstic springs were much more active than at present. The Antalya tufa is composed of pure low-Mg calcite and this is typical of cool-water tufas in the Mediterranean region. Tectonic uplift of the tufa mass created an upper 300 m terrace and glacio-eustatic sea-level change in the Early to Middle Pleistocene created a lower terrace (c.100– 200 m a.s.l.). Subsequent fluvial activity has cut deep gorges in the tufa and the coastal sections have been eroded by wave action (Glover and Robertson 2003). Thus, the present landscape is the product of complex interactions between a wide range of karstic and nonkarstic geomorphological processes and materials over the last two million years or so. In many places the surface of the tufa is crusted and displays complex secondary karst forms following dissolution and recementation of the tufa deposits (Glover and Robertson 2003). In many areas deep terra rossa type soils have developed. The deposition of secondary carbonates may form cemented alluvial gravels. These are widespread in Mediterranean karst environments and the

306

John Lewin and Jamie Woodward

(a)

(b)

Fig. 10.14. Two limestone gorges at different stages of development in the Mediterranean region. Both are the product of fluvial incision and subaerial weathering of the bedrock cliffs. (a) The Moracha Canyon in Montenegro and (b) the Vikos Gorge in north-west Greece (photos: Jamie Woodward).

best-developed records are typically Pleistocene in age (Vita-Finzi 1969; Hamlin et al. 2000; Macklin et al. 2002; Woodward et al. 2008). This can help preserve the alluvial record and has allowed the development of long-term records of Pleistocene river behaviour through the application of uranium-series dating methods (Chapter 11). Figure 10.16b shows a section in over 20 m of cemented Pleistocene alluvial gravels in one of the wider reaches of the Moracha Canyon in Montenegro to the north of Podgorica (Figure 10.1). Ambient temperature deposited carbonates in streams have led to the formation of tufa barrages, cascades, or sheets. These can be impressive and deposition can take place above growing barriers, as in the case of the River Krka in Croatia where there is an alluvial plain above the 20-m high Topolje Falls (Jennings 1985), whilst the

River Korana is divided into sixteen separate lakes along a 12-km gorge at Plitvice (Figure 10.1). Thermogene travertines are developed, again spectacularly, in Turkey at Pamukkale, at the Roman city of Hierapolis where 35–39◦ C waters emerge at the surface to form a very distinctive topography (Figure 10.16c). The thick travertines at Bagni de Tivoli near Rome have been quarried for over 2,000 years. Travertine, along with marble, was widely used as a building material in Classical Rome because it is strong yet relatively light due to its open texture. These carbonate rocks represent the foundation and essence of classical architecture. Pentecost (1995) provides a useful catalogue of both ambient-temperature and thermogene deposits in Europe and Turkey, whilst sites in Mediterranean North Africa are discussed by Ford and Pedley (1996).

Karst Geomorphology

307

Fig. 10.15. Relict karst pinnacles in the White Desert of Egypt (photo: Tony Waltham).

Karst Systems and Environmental Change Given the marked environmental changes that took place in the Mediterranean region during the Pleistocene and Holocene (Chapters 2, 4, 9, 11, and 12), it is clear that karst morphogenesis has been considerably affected by factors other than the autogenic effects so far discussed as, for example, in the cave evolution model presented in Figure 10.7. In broad terms, it is possible to identify four major environmental changes that have impacted upon the karst systems of the region. First, the elevation and exhumation of limestone surfaces, river basin evolution, and karst development have proceeded alongside continuing tectonic activity in the Mesozoic and Cenozoic eras (e.g. Harvey and Wells 1987; Chapter 1) and, initially, under a climate warmer than present (Chapter 4). In southern France, bauxite deposits (named from the cave-ridden spur of Les Baux) were created by karst weathering even before submergence in a Late Cretaceous sea (Ager et al. 1980). Second, the availability of relief has varied, both in response to eustatic Quaternary sea level change (of the

order of 120 m) and earlier in response to the remarkable Messinian (Late Miocene) drying of the Mediterranean Sea basin (Hsü 1972; Adams et al. 1977), when regional base levels fell by more than 1,000 m (Chapter 1). Limestone areas that are now coastal, or near to sea level, were elevated above the littoral zone during the Messinian Salinity Crisis and, repeatedly, during Pleistocene cold-stage low sea-level stands. This means that karst forms are now ‘drowned’ deep in the vadose zone. Recently, Bakalowicz et al. (2003) have considered the development of the deep karst aquifers in the Mediterranean coastal zone that formed during the Late Miocene Messinian Salinity Crisis around 5.5 Ma. The sea-level fall during this period led to the development of large canyons in the lower reaches of the major river systems. This base level fall also led to the formation of very deep karst systems that lie well below present sea levels. The residence time of the groundwaters stored in these deep coastal karsts is not known. These deep karst aquifers have very long vertical conduits and are unique to the Mediterranean region (Mocochain et al. 2006; Bakalowicz et al. 2008). From a resource management perspective, it is important to establish

(a) (b)

(c)

Karst Geomorphology

309

Fig. 10.17. Karst terrain in north-east Majorca showing bare limestone slopes and thick terra rossa soils on the valley floor (photo: Jamie Woodward).

their water resource potential and their role in the contemporary hydrological cycle. Bakalowicz et al. (2003) have discussed the hydrogeological significance of these deep aquifer systems and the difficulties involved in monitoring submarine karst springs and the dynamics of freshwater and sea water exchange in the deep phreatic zone. In Dalmatia these springs are known as vrulje and they have been reported from around the Mediterranean coastal karst. Fleury et al. (2008) monitored the behaviour of two submarine outlets from a karstic system in southern Spain during the 1999–2000 hydrological year. One of the conduits was highly responsive to rapid infiltration across the terrestrial catchment during highintensity rainfall events. Fig. 10.16. (opposite) Three landscape features in the Mediterranean produced by the precipitation of secondary carbonates. (a) The coastal cliffs in tufa deposits at Anatalya on the southern coast of Turkey. (b) A 20-m exposure in cemented alluvial gravels of Pleistocene age in the lower reaches of the Moracha River in Montenegro. (c) The dramatic pools and slopes at Pamukkale in south-west Turkey. All locations are shown on Figure 10.1. (Photos: (a) and (c), Tony Waltham; (b), Jamie Woodward).

Third, the effects of Pleistocene climate fluctuations have been superimposed upon the broad framework of tectonic activity across the region to impose, in very general terms, a series of wet/dry and cold/warm phases on karstic systems (Chapters 2, 4, 11, and 12). These shifts in Quaternary climate altered the balance between physical and chemical weathering in limestone areas, with more intense physical weathering (and in places more extensive cirque and valley glaciation) of exposed upland limestones during cold episodes (Chapter 12). Thus cliff weathering in many limestone gorges produced Pleistocene talus slopes and boulder-bed streams in excess of those generated by cliff collapse and rockfall today. Lithological and mineralogical analysis of Pleistocene alluvial materials suggests much greater incorporation of physical weathering products in cold phase deposits than later (Lewin et al. 1991; Woodward et al. 1992). The Quaternary also involved periods of higher and lower water tables, so that valley networks and river reaches, and cave systems that are now dry were being actively developed by flowing water. Groundwater bodies beneath some of the desert areas of North Africa were

310

John Lewin and Jamie Woodward

TABLE 10.3. Characteristics of active karst settings and passive karst settings for rockshelter and cave entrance environments in limestone terrains Active Karst Setting r Linked to an internal cavern or conduit system r Dripping vadose waters r Seasonal water flows and ponding r Range of hydrological pathways r Precipitation of calcite and other minerals r Inwashing of fine sediments via conduits in the host bedrock r Development of vegetation within the site r Mineralization of macroscopic plant remains r Strong chemical diagenesis and mineral alteration r Humidity may encourage host rock breakdown by frost action r Evidence of erosion and sediment removal by invasive karst waters Passive Karst Setting r No significant links with an internal conduit system r Dry site without flowing or dripping water r Limited inwashing of sediments via karstic cavities r Highly localized or no chemical precipitation r Limited vegetation growth in the site r Desiccation of macroscopic plant remains r Limited chemical diagenesis and mineral alteration r Import of fine sediments through the shelter opening may be dominant r Limited host rock weathering by solution r Subaerial processes are dominant Note: See Fig. 10.8. Sites may shift between these two end-members in response to climate change. Source: Modified from Woodward and Goldberg (2001).

developed under more humid Pleistocene conditions, so that they now constitute a fragile, fossil resource that is not effectively being replenished. Fourth, the removal of a soil and vegetation cover, particularly following human activity during the course of the Holocene (Chapters 6 and 9), has transformed karst systems. Thus deep weathering forms developed under a soil cover (known as rundkarren and generally of rounded form with residual core stones) have become exposed on the surface to form small-scale landscapes of pinnacles, blocks, and chasms as shown in the karst terrain of central Majorca in Figure 10.17. These terrains may be modified by the development of sharp-edged flutes (rillenkarren) and channels (rinnenkarren) produced subsequently by the effects of subaerial solution on bare limestone surfaces. Thus some karst features may be composite forms that owe their origin to a combination of different climatic episodes (and associated process regimes involving exhumation and/or burial by Quaternary sediments) just as cave systems may develop under different phases of water flow and chemical balance (Figures 10.7 and 10.8). Complex surface topography, with a combination of relict and active forms, may have an exposed local relief of

the order of 10 m, as in the ‘roches ruiniformes’ of the French Causses or the ‘cuidad encantada’ of the Sierra delibar in southern Spain.

Archives of Environmental Change Sediments in Limestone Rockshelters and Cave-mouth Environments Caves and rockshelters are found wherever hard limestones are present in the Mediterranean basin. Most deep caves owe their formation to karstic processes to a greater or lesser extent, while some shallow rockshelters may result from non-karstic processes such as fluvial erosion and undercutting, or physical weathering of a gorge wall, for example. In the same way the sediments that accumulate in cave-mouth and rockshelter settings will commonly represent some combination of karstic and non-karstic processes—although, in practice, the division is not always clear cut. Woodward and Goldberg (2001) have used the geomorphological and hydrological characteristics of cave-mouth and rockshelter environments to describe active karst settings and passive karst settings (Table 10.3), because the local context is a very important control on the nature of the accumulating sediment. Caves and rockshelters in the Mediterranean have provided shelter for both animals and humans throughout much of the Quaternary and they represent a major archaeological resource. Most of the work on the clastic sedimentary records in rockshelter and cave-mouth environments has been carried out as part of archaeological excavations of Palaeolithic and Mesolithic records (ibid.). Figure 10.18 illustrates the wide range of processes that can transport fine-grained sediments to rockshelter and cave mouth environments in karst settings (Woodward and Bailey 2000). It is also important to appreciate, however, that these processes

Fig. 10.18. (opposite) Rockshelter and cave entrance environments can form important sediment sinks and they represent a major archaeological resource. (a) The opening of Asprochaliko rockshelter in the Louros Valley of north-west Greece. This site contains evidence of Middle and Upper Palaeolithic occupation and was excavated by Eric Higgs and his team in the 1960s (photo: Jamie Woodward). (b) Schematic cross-section of a Mediterranean limestone rockshelter showing the processes that can deliver fine sediments to the site and the natural and archaeological materials that can be used for radiometric dating (based on Woodward and Bailey 2000, and Schwarz and Rink 2001). The allogenic (external) sediment sources are shown in white boxes and the autogenic (internal) ones are shown in the shaded box. Compare to Figure 10.8 and Table 10.3.

(a)

(b)

Detrital component in bedrock

Fine sediments washed through joints, bedding planes and larger conduits.

Physical and chemical weathering of the host limestone bedrock. Debris from tool manufacture and other activities.

Colluvial inputs including fine sediments washed down the gorge wall during storm events, wash from adjacent slopes and mass movement.

Aeolian inputs such as far-travelled dust, tephra and local silts and sands.

Older bedrock profiles

Stalactites ‘Soda straw’ stalactites

Human imports of fine sediment on wet feet and animal carcasses. Raw materials for stone tool manufacture.

Windblown detritus

Bones and teeth Flowstone

Fireplace Eroded channel

Weathered roof block

Natural erosion surface

Fallen stalagmite

Sink hole

Old eroded remnant

Direct fluvial input of fine slackwater sediments during large floods. Fine sediment deposition during high lake or sea level stands.

312

John Lewin and Jamie Woodward

can also be destructive and may remove parts of the sedimentary record. Gaps in the record may also represent periods of non-deposition. In the case of Franchthi Cave in southern Greece, for example, Farrand (2000) points out that at least half the interval between c.26,850 and 6,000 cal. years BP is not represented by any deposits in the excavated parts of the site. Thus, detailed dating programmes are needed to establish the timeframe and chronological resolution of the sedimentary record (Bailey and Woodward 1997; Schwartz and Rink 2001). Despite the poor stratigraphic resolution of some sites in comparison to other depositional environments, limestone caves and rockshelters commonly contain a range of materials that can be dated (Figure 10.18). These include materials produced by human activity (e.g. charcoal fragments and burnt flints in hearths), karstic processes (speleothems), and allogenic sediments introduced by natural processes (e.g. tephras and aeolian sands). All these materials can be dated using modern scientific methods and these are summarized in Figure 10.18. The use of large-format thin sections to study the microstratigraphy of these deposits has allowed detailed reconstructions of site formation processes (e.g. Goldberg and Bar-Yosef 1998) and recent work on sediment source identification has allowed these records to be placed in their broader geomorphological context (e.g. Woodward et al. 2001).

Speleothems and Quaternary Palaeoclimates Recent work on speleothems from karstic caves in Israel has provided excellent high resolution records of climate change for the eastern Mediterranean region (BarMatthews et al. 1999, 2000, 2003). Soreq Cave is located west of Jerusalem and around 40 km from the Mediterranean Sea (Figure 10.1). It lies close to the modern 500 mm isohyet at about 400 m above sea level. The speleothems from Soreq Cave and from Peqiin Cave in the north of Israel have yielded a continuous and high resolution record of karstic processes and climate change spanning the last 250,000 years (Figure 10.19). The carbonate materials have been dated using high precision uranium-series methods. These records match closely with the marine oxygen isotope record and have demonstrated that many features of the climate record for the last cold stage in the North Atlantic (such as Heinrich Events) also impacted upon the eastern Mediterranean region (Bar-Matthews et al. 1999) (Chapters 2 and 4).

The caverns in these semi-arid karst landscapes are highly sensitive environmental systems that are well coupled to changes in temperature and precipitation— although it is important to point out that a full understanding of the contemporary karst hydrology is needed before the palaeoclimatic record can be fully appreciated (ibid.). A very significant feature of recent work on the speleothem records in Israel has been the exploration of linkages between the high resolution speleothem data and the marine records from the eastern Mediterranean Sea (Bar-Matthews et al. 2003). This research has shown that some periods of sapropel formation were associated with enhanced precipitation across the eastern Mediterranean region. Sapropels are dark, organic-rich horizons that formed periodically on the bottom of the Mediterranean Sea, and are discussed in Chapter 2. Information on Pleistocene and Holocene sea-level change can also be obtained from speleothems where they have developed in caverns in coastal karst systems that are submerged during high sea-level stands (Chapter 13).When these caverns become flooded by sea water, speleothem growth is halted and they become covered with marine biogenic overgrowths (Suri´c et al. 2005). These overgrowths can be dated by radiocarbon to provide geochronologies for the timing of cavern flooding by sea water although the precise timing of inundation may be conditioned by the geomorphological setting and local karst hydrology. At some sites flooding is by fresh groundwater as local water tables rise in accordance with sea-level rise. Thus, cavern shape and elevation, groundwater conduit gradients, and cavern distance from the coast are important factors in site selection. Suri´c et al. (2005) present data from three submerged caves along the Croatian coast where eight speleothems (ranging from 38.5 to 17 m below present mean sea level) were sampled. Speleothem deposition took place in these caves during the last cold stage between c.37,000 and 22,000 years BP (when global sea levels were around 120 m lower than present). These workers have produced a sea-level curve for the Late Pleistocene and Holocene that is in broad agreement with work on the Tyrrhenian coast of Italy and the Mediterranean coast of France. Similarly, Bard et al. (2002) have estimated interglacial (MIS 7) sea levels from speleothems preserved in Argentarola Cave in Italy.

The Conservation of Karst Environments Today, worldwide academic interest in these karst environments must also be coupled with an appreciation of the value and fragility of Mediterranean karst

Karst Geomorphology

313

–9 1

2

3

4

5

–8

d18O % (PDB)

–7

–6

–5

–4

–3 S1

S3

S4

S5

–2 0

20

40

60 80 Age (thousand years)

100

120

140

Fig. 10.19. A high-resolution oxygen isotope record (‰18 O) from speleothems in Soreq Cave, Israel, spanning the last 140,000 years (modified from Bar-Matthews et al. 1999). This record shows how the regional climate has fluctuated between warmer and wetter periods (peaks) to cooler and drier periods (troughs). It also shows that climate change can be abrupt (see Chapters 4, 9, and 11). Marine oxygen isotope stages 1–5 are shown along with sapropels S1, S3, S4, and S5. Sapropel formation in the eastern Mediterranean Sea is discussed in detail in Chapter 2.

landscapes and groundwater resources (Plagnes and Bakalowicz 2002; Bakalowicz et al. 2008). The conservational aspects of limestone and karsts in the region is attracting increasing attention (Gams et al. 1993). This includes threats to groundwater quality by pollution (both point and diffuse sources and waste disposal) and the potentially damaging exploitation of fossil groundwaters, the destruction of valued landscapes and archaeological sites by quarrying, and by urban/recreational development and land use pressure (Figure 10.20). Karstic lakes are often ecologically important sites but are susceptible to eutrophication and habitat degradation (Reed et al. 2008). There are numerous examples of the unwise use of subsurface water systems which

have not been adequately understood, but which have suffered from pollution or resource depletion. Action to protect aquatic environments by the European Commission has involved vulnerability and risk assessment of carbonate (karst) aquifers. The exploitation of limestone terrains is nothing new—as shown by the huge Carrara marble quarries in Italy that were a resource for Renaissance art and architecture—but the scale of recent transformations is unprecedented. Population pressure and the large number of tourists visiting underground cave systems can also threaten cave environments (including their atmosphere, flora, and fauna) without access management. Limestone caves and rockshelters also

314

John Lewin and Jamie Woodward

Human activities

Impacts on Mediterranean karsts

Effects Karst terrain

Adjacent non-karst terrain Deforestation

Impoverished ecology

Loss of biota Reduced evapotranspiration

Agriculture Increased runoff and erosion Urbanization and industry

Soil degradation and erosion Sedimentation of caves

Increased sediment discharge Water quality deterioration Acid rain

Quarrying and mining

Landform destruction Waste water discharge

Tourism and recreation

KARST ECOSYSTEM DEGRADATION

Rock and mineral removal

Chemical wastes

Military activities

Soil collapse Wear and tear Water exploitation Reduced allogenic recharge

Spring desiccation

– dams (upstream) – groundwater pumping – dams (downstream)

Water-table lowering

Sea water intrusion

Inundation

Drowned karst systems

Fig. 10.20. Human activities in the Mediterranean region and their potential impact on non-karst and karst terrains (modified from Williams 1993).

TABLE 10.4. Karst sites in the Mediterranean region with World Heritage status Country Croatia France and Spain Slovenia Spain Turkey Montenegro

Karst environment Plitvice Lakes National Park Pyrenees-Mount Perdu Škocjan Caves Altimera Cave Atapuerca Cave Pamukkale Durmitor National Park

Designation Natural Heritage Natural and Cultural Natural Heritage Cultural Heritage Cultural Heritage Natural and Cultural Natural Heritage

Source: Modified from Hamilton-Smith (2004).

preserve valuable records of Quaternary environmental change that include a rich archaeological resource on the history of Palaeolithic and Mesolithic cultures in particular. Limestone caves in the Mediterranean region— including Altimira Cave in Spain and Chauvet in France (Figure 10.1)—contain some of the best-known examples of Upper Palaeolithic cave art and these also require

careful conservation and management. A recent and very positive development has been the designation of World Heritage Site status by UNESCO for several karst and cave sites in the Mediterranean region and these are listed in Table 10.4.

Conclusions Karst terrains are a very significant component of the physical geography of the Mediterranean basin because the effects of both carbonate dissolution and precipitation processes are widespread. It can be argued that carbonate rocks are an important unifying feature of much of the Mediterranean basin. Many of the key mountain ranges such as the Pindus Mountains of Greece, the Apennines of Italy, and the Dinaric Alps of the former Yugoslavia are dominated by uplifted limestone terrains and these landscapes include glaciokarsts, deep

Karst Geomorphology

limestone gorges, large dolines, sinking streams, and extensive cave systems. Limestone bedrock coasts are also a key feature of the Mediterranean environment (Chapter 13). Much of the river water that flows into the Mediterranean Sea has passed through karst terrain, so this environment is an important influence on both river regimes and water quality (Chapter 8). Karst landscapes pose particular engineering problems and ground collapse can be a significant geohazard in many areas. Karst also leads to the formation of distinctive red terra rossa soils (Chapter 6) and vegetation communities and it has influenced the development of agricultural practices and water resource management strategies for thousands of years. In some parts of the Mediterranean region—particularly those in North Africa—the karsts have yet to be fully explored. This is reflected in the much smaller literature on the karsts of several countries including Libya, Algeria, and Tunisia, for example (see entries in Gunn 2004). An understanding of karstic processes is essential for many disciplines and may provide insights in unexpected contexts. For example, recent work at the Early Bronze Age fortified site of Vayia, in southern Greece, has shown that the variable development and orientation of karren features on limestone blocks found within cairns could be used to establish a relative dating framework for architectural remains and to establish that the cairns themselves were formed in antiquity (Tartaron et al. 2006). It is also important to appreciate the significance of carbonate precipitation dynamics in the karst systems of the region. The formation of tufa, travertine, speleothems, and calcrete is also a key feature of Mediterranean karst environments and this has produced a highly distinctive suite of karstic features and landscapes. The Mediterranean basin contains some of the most extensive and spectacular Pliocene- and Quaternary-age tufa deposits in the world (Glover and Robertson 2003) including the distinctive terrains at Pamukkale and Plitvice. Carbonate precipitates often constitute important archives of Quaternary environmental change that can be dated by uranium-series methods. This approach has provided new insights into the Pleistocene glacial record in the Mediterranean and recent work on speleothems in Israel, in particular, has produced very detailed and globally significant records of environmental change for the last 250,000 years (e.g. Bar-Matthews et al. 2003). Mediterranean karst landscapes have been shaped by a range of spatially and temporally variable environmental fluctuations in the recent geological past. In broad terms these reflect the variable impacts of long-term tectonic history and Quaternary climate change and, in the Holocene, the

315

increasing intensity of human activity across the basin from Neolithic to modern times.

References Adams, C. G., Benson, R. H., Kidd, R. B., Ryan, W. B. F., and Wright, R. C. (1977), The Messinian salinity crisis and evidence of late Miocene eustatic changes in the world ocean. Nature 269: 383–6. Ager, D. V. (1980), The Geology of Europe. McGraw Hill, London. Albritton, C. C., Brooks, J. E., Issawi, B., and Swedor, A. (1990), Origin of the Qattara Depression, Egypt. Bulletin of the Geological Society of America 102: 952–60. Bailey, G. N. and Woodward, J. C. (1997), The Klithi deposits: sedimentology, stratigraphy and chronology, in G. N. Bailey (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, i. Excavation and Intra-site Analysis at Klithi. Cambridge, McDonald Institute for Archaeological Research, 61–94. Turner, C., Woodward, J. C., Macklin, M. G., and Lewin, J. (1997), The Voidomatis basin: an introduction, in G. N. Bailey (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, i. Klithi in its Local and Regional Setting. Cambridge, McDonald Institute for Archaeological Research, 321–45. Bakalowicz, M., Fleury, P., Jouvencel, B., Promé, J. J., Becker, P., Carlin, T., Dörfliger, N., Seidel, J. L., and Sergent, P. (2003), Coastal karst aquifers in Mediterranean regions: a methodology for exploring, exploiting and monitoring sub-marine springs. Tecnologia de la Intrusion de Agua de Mar en Acuiferos Costeros: Paises Mediterraneos. IGME, Madrid. El Hakim, M., and El-Hajj, A. (2008), Karst groundwater resources in the countries of eastern Mediterranean: the example of Lebanon. Environmental Geology 54: 597–604. Bard, E., Antonioli, F., and Silenzi, S. (2002), Sea-level during the penultimate interglacial period based on a submerged stalagmite from Argentarola Cave (Italy). Earth and Planetary Science Letters 196: 135–46. Bar-Matthews, M., Ayalon, A., Kaufman, A., and Wasserburg, G. J. (1999), The Eastern Mediterranean palaeoclimate as a reflection of regional events: Soreq cave, Israel. Earth and Planetary Science Letters 166: 85–95. (2000), Timing and hydrological conditions of Sapropel events in the Eastern Mediterranean, as evident from speleothems, Soreq cave, Israel. Chemical Geology 169: 145–56. Ayalon, A., Gilmour, M., Matthews, A., and Hawksesworth, C. (2003), Sea-land oxygen isotopic relationships from planktonic foraminifera and speleothems in the Eastern Mediterranean region and their implication for paleorainfall during interglacial intervals, Geochimica et Cosmochimica Acta, 67: 3181–99. Beckinsale, R. P. and Chorley, R. J. (1991), The History of the Study of Landforms or the Development of Geomorphology 1890 to 1950. New York, Routledge. Biju-Duval, B., Dercourt, V., and Le Pichon, X. (1977), From the Tethys ocean to the Mediterranean sea: a plate tectonic model of the evolution of the western Alpine system, in B. Biju-Duval and L. Montadert (eds.), The Structural History of the Mediterranean Basins. Éditions Technip, Paris, 143–64.

316

John Lewin and Jamie Woodward

Bonacci, O., Ljubenkov, I., and Roje-Bonacci, T. (2006), Karst flash floods: an example from the Dinaric karst (Croatia). Natural Hazards and Earth Systems Science 6: 195–203. Brewster, H. (1997), The River Gods of Greece: Myths and Mountain Waters in the Hellenic World. I. B. Tauris, London. Calaforra, J. M. and Pulido-Bosch, A. (2003), Evolution of the gypsum karst of Sorbas (SE Spain). Geomorphology 50: 173–80. Courbon, P. and Chabert, C. (1986), Atlas des Grandes Cavities Mondiales. UIS-FFS. Cviji´c, J. (1893), Das Karstphänomen. Geographische Abhandlungen herausgegeben von A. Penck [Geographical Proceedings. Published by A. Penck] 5/3: 218–329. Dewey, J. F., Pitman III, W. C., Ryan, W. B. F., and Bonnin, J. (1973), Plate tectonics and the evolution of the Alpine system. Bulletin of the Geological Society of America 84: 3137–80. Farrand, W. R. (2000), Depositional History of Franchthi Cave: Sediments, Stratigraphy and Chronology: Fascicle 12. Indiana University Press, Bloomington, Ind. Ferrarese, F., Sauro, U., and Tonello, C. (1998), The Montello Plateau. Zeitschrift für Geomorphologie, Suppl.109: 41–62. Fleury, P., Bakalowicz, M., de Marsily, G., and Cortes, J. M. (2008), Functioning of a coastal karstic system with a submarine outlet, in southern Spain. Hydrogeology Journal 16: 75–85. Ford, D. C. (2004), Karst in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 473–5. and Pedley, H. M. (1996), A review of tufa and travertine deposits of the world. Earth Science Reviews 41: 117–75. and Williams, P. W. (1989), Karst Geomorphology and Hydrology. Unwin Hyman, London. Frumkin, A. (1994), Hydrology and denudation rates of halite karst. Journal of Hydrology 162: 171–89. Gale, S. J. and Hoare, P. G. (1997), The glacial history of the northwest Picos de Europa of northern Spain. Zeitschrift für Geomorphologie, NS 41: 81–96. Gilbertson, D. D., Hoare, P. G., Hunt, C. O., Jenkinson, R. D., Lamble, A. P., O’Toole, C., van der Veen, M., and Yates, G. (1993), Late Holocene environmental change in the Libyan predesert. Journal of Arid Environments 24: 1–19. Gams, I. (1978), The Polje: The problems of its definition. Zeitschrift für Geomorphologie 22: 170–81. Nicod, J., Julian, M., Anthony, E., and Sauro, U. (1993), Environmental change and human impacts on the Mediterranean karsts of France, Italy and the Dinaric region, in P. W. Williams (ed.), Karst Terrains, Environmental Changes, Human Impact. Catena Suppl. 25: 59–98. Gillieson, D. (1996), Caves: Processes, Development and Management. Blackwell, Oxford. Glover, C. and Robertson, A. H. F. (2003), Origin of tufa (coolwater carbonate) and related terraces in the Antalya area, SW Turkey. Geological Journal 38: 329–58. Goldberg, P. and Bar-Yosef, O. (1998), Site formation processes in Kebara and Hayonim Caves and their significance in Levantine prehistoric caves, in T. Akazawa, K. Aoki, and O. Bar-Yosef (eds.), Neandertals and Modern Humans in Western Asia. Plenum, New York, 107–23. Gunn, J. (2004), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London. Hamilton-Smith, E. (2004), World Heritage Sites, in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 777–9. Hamlin, R. H. B., Woodward, J. C., Black, S., and Macklin, M. G. (2000), Sediment fingerprinting as a tool for interpreting long-

term river activity: the Voidomatis basin, NW Greece, in I. D. L. Foster (ed.), Tracers in Geomorphology. John Wiley & Sons, Chichester, 473–501. Harvey, A. M. and Wells, S. G. (1987), Response of Quaternary fluvial systems to differential epeirogenic uplift: Aguas and Feos river systems, southeast Spain. Geology 15: 689–93. Herak, M. and Stringfield, V. T. (eds.) (1972), Karst: Important Karst Regions of the Northern Hemisphere. Elsevier, Amsterdam, 551. Hodge, E. J., Richards, D. A., Smart, P. L., Andreo, B., Hoffmann, D. L., Mattey, D. P., and González-Ramón, A. (2008), Effective precipitation in southern Spain (266 to 46 ka) based on a speleothem stable carbon isotope record. Quaternary Research 69: 447–57. Hsü, K. J. (1972), When the Mediterranean dried up. Scientific American 227: 27–36. Hughes, P. D., Woodward, J. C., Gibbard, P. L., Macklin, M. G., Gilmour, M. A., and Smith, G. R. (2006), The glacial history of the Pindus Mountains, Greece. Journal of Geology 114: 413–34. Jakucs, L. (1977), Morphogenetics of karst regions. Unwin Hyman, London. Jennings, J. N. (1985), Karst Geomorphology. Blackwell, Oxford. Jourde, H., Roesch, A., Guinot, V., and Bailly-Comte, V. (2007), Dynamics and contribution of karst groundwater to surface flow during Mediterranean flood. Environmental Geology 51: 725–30. King, G., Sturdy, D., and Bailey, G. N. (1997), The tectonic background to the Epirus landscape, in G. N. Bailey (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece. McDonald Institute, Cambridge, ii. 541–58. Klimchouk, A. (2004), Morphometry of Caves, in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 524–6. Lowe, D., Cooper, A., and Sauro, U. (eds.) (1996), Gypsum Karst of the World. International Journal of Speleology (Special Issue) 25: 3–4. Lewin, J., Macklin, M. G., and Woodward, J. C. (1991), Late Quaternary fluvial sedimentation in the Voidomatis Basin, Epirus, northwest Greece. Quaternary Research 35: 103–15. Lowry, D. C. and Jennings, J. N. (1974), The Nullabor karst Australia. Z. Geom. 18: 35–81. Macklin, M. G., Fuller, I. C., Lewin, J., Maas, G. S., Passmore, D. G., Rose, J., Woodward, J. C., Black, S., Hamlin, R. H. B., and Rowan, J. S. (2002), Correlation of Late and Middle Pleistocene fluvial sequences in the Mediterranean and their relationship to climate change. Quaternary Science Reviews 21/14/15: 1633–44. Milanovi´c, P. T. (2000), Geological Engineering in Karst. Belgrade, Zebra. Mocochain, L., Clauzon, G., Bigot, J., and Brunet, P. (2006), Geodynamic evolution of the peri-Mediterranean karst during the Messinian and the Pliocene: evidence from the Ardèche and Rhône Valley systems canyons, Southern France. Sedimentary Geology 188/9: 219–33. Palmer, A. N. (1991), Origin and morphology of limestone caves. Bulletin of the Geological Society of America 103: 1–21. Pena, J. L., Sancho, C., and Lozano, M. V. (2000), Climatic and tectonic significance of Late Pleistocene and Holocene tufa deposits in the Mijares River canyon, eastern Iberian Range, Northeast Spain. Earth Surface Processes and Landforms 25: 1403–17.

Karst Geomorphology Pentecost, A. (1995), The Quaternary travertine deposits of Europe and Asia Minor. Quaternary Science Reviews 14: 1005– 28. Perritaz, L. (2004), Africa, North in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 13–16. and Monbaron, M. (1998), Geomorphological approach to the Aït Abdi Karst Plateau (Central High Atlas, Morocco). Zeitschrift für Geomorphologie 109: 83–104. Plagnes, V. and Bakalowicz, M. (2002), The protection of a karst water resource from the example of the Larzac karst plateau (south of France): a matter of regulations or a matter of process knowledge? Engineering Geology 65: 107–16. Reed, J. M., Leng, M. L., Ryan, S., Black, S., Altinsaçli, S., and Griffiths, H. I. (2008), Recent habitat degradation in karstic Lake Uluabat, western Turkey: A coupled limnological– palaeolimnological approach. Biological Conservation 141: 2765–83. Rossi, C. (2004), Picos de Europa, Spain in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 582–5. Schwarz, H. P. and Rink, W. J. (2001), Dating methods for sediments of caves and rockshelters. Geoarchaeology 16, 355–71. Smart, C. C. and Worthington, S. R. H. (2004), Springs, in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 699–703. Smart, P. L. (1986), Origin and development of glacio-karst closed depressions in the Picos de Europa, Spain. Zeitschrift für Geomorphologie, NS 30; 423–43. Smith, B. J., Warke, P. A., and Moses, C. A. (2000), Limestone weathering in contemporary arid environments: a case study from southern Tunisia. Earth Surface Processes and Landforms 25: 1343–54. Suri´c, M., Juraˇci´c, M., Horvatinˇci´c, N., and Broni´c, I. K. (2005), Late Pleistocene–Holocene sea level rise and the pattern of coastal karst inundation: records from submerged speleothems along the Eastern Adriatic Coast (Croatia). Marine Geology 214: 163–75. Tartaron, T. F., Pullen, D. J., and Noller, J. S. (2006), Rillenkarren at Vayia: geomorphology and a new class of Early Bronze Age fortified settlement in Southern Greece. Antiquity 80: 145–60. Van Houten, F. B. (1980), Latest Jurassic-Early Cretaceous regressive facies, northeast Africa Craton. American Association of Petroleum Geologists Bulletin 64: 857–67. Vita-Finzi, C. (1969), The Mediterranean Valleys: Geological Changes in Historical Times. Cambridge, Cambridge University Press. Waltham, A. C. (1978), The caves and karst of Astraka, Greece. Transacations of the British Cave Research Association 5: 1–12.

317

(2001), Pinnacles and barchans in the Egyptian desert. Geology Today 17: 101–4. White, W. B. (1988), Geomorphology and Hydrology of Karst Terrains. Oxford University Press, Oxford. Williams, P. W. (1993), Environmental change and human impact on karst terrains: an introduction, in P. W. Williams (ed.), Karst Terrains, Environmental Changes, Human Impact. Catena Suppl. 25: 1–19. (2004), Karst Evolution, in J. Gunn (ed.), Encyclopedia of Caves and Karst Science. Fitzroy Dearborn, London, 475–8. Windley, B. F. (1984), The Evolving Continents. John Wiley & Sons, Chichester. Woodward, J. C. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water Quality in River Catchments. John Wiley and Sons, Chichester, 365–89. (1997), Late Pleistocene rockshelter sedimentation at Megalakkos, in G. N. Bailey (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, ii. Klithi in its Local and Regional Setting. McDonald Institute for Archaeological Research, Cambridge, 377–93. and Bailey, G. N. (2000), Terminal Pleistocene sediment sources and geomorphological processes recorded in rockshelter sequences in northwest Greece in I. D. L. Foster (ed.), Tracers in Geomorphology. John Wiley & Sons, Chichester, 521–51. and Goldberg, P. (2001), The sedimentary records in Mediterranean rockshelters and caves: archives of environmental change. Geoarchaeology: An International Journal 16/4: 327–54. Lewin, J., and Macklin, M. G. (1992), Alluvial sediment sources in a glaciated catchment: the Voidomatis basin, northwest Greece. Earth Surface Processes and Landforms 17/3: 205–16. Hamlin, R. H. B., Macklin, M. G., Karkanas, P., and Kotjabopoulou, E. (2001), Quantitative sourcing of slackwater deposits at Boila Rockshelter: A record of Lateglacial flooding and Palaeolithic settlement in the Pindus Mountains, Northwest Greece. Geoarchaeology: An International Journal 16/5: 501–36. Macklin, M. G., and Smith, G. R. (2004), Pleistocene glaciation in the mountains of Greece, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations—Extent and Chronology, i. Europe Elsevier, Amsterdam, 155–73. Hamlin, R. H. B., Macklin, M. G., Hughes, P. D., and Lewin, J. (2008), Glacial activity and catchment dynamics in northwest Greece: Long-term river behaviour and the slackwater sediment record for the last glacial to interglacial transition. Geomorphology 101: 44–67.

This chapter should be cited as follows Lewin, J. and Woodward, J. C. (2009), Karst geomorphology and environmental change, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 287–317.

This page intentionally left blank

11

River Systems and Environmental Change Mark Macklin and Jamie Woodward

Introduction Linking river behaviour and drainage basin evolution to Quaternary environmental change, most notably the effects of climatic variability, tectonics, and human activity on runoff and sediment delivery, has a long history of research in the Mediterranean areas of Europe, North Africa, and the Near East. This field of research was initially stimulated by the (re)discovery at the beginning of the twentieth century of many Classical Period remains buried by river alluvium; perhaps the best known of which is the site of Olympia in western Greece (Huntington 1910) (Figure 11.1). The widespread evidence for large-scale shifts in river channel positions and the rapid growth of deltas and coastal alluvial plains in historical times (Judson 1963; Raphael 1973; Kraft et al. 1980; and Chapter 13) also provided much impetus for this research. In addition, archaeological investigations carried out soon after the Second World War in Algeria (Gaucher 1947), Italy (Selli 1962), Libya (McBurney and Hey 1955) and Spain (Gigout 1959) resulted in the recovery of large numbers of Palaeolithic stone tools from Pleistocene fluvial deposits. These early examples of what has now become more widely known as ‘geoarchaeology’ (Davidson and Shackley 1976; Butzer 1977) or ‘alluvial archaeology’ (Macklin and Needham 1992) were, with their strong interdisciplinary focus, highly innovative and ahead of their time in the way they integrated archaeology, geomorphology, and geochronology. Building on this theme, the principal aim of this chapter is to consider how river systems in the Mediterranean

region have responded to the environmental changes that took place during the Late Quaternary–a time interval corresponding approximately to the last 130,000 years. There are a number of reasons for choosing this period for reviewing river-environment interactions in the Mediterranean: 1. It encompasses the last glacial–interglacial cycle (c.130 to 10 ka) for which there is now abundant global evidence from polar ice cores, speleothem records, and lake and marine sediments, for both longand short-term changes in climate. These changes included massive reorganizations of the atmosphereocean-cryosphere systems—often over timescales of less than 100 years (Lowe and Walker 1997)—and they are clearly recorded in the Mediterranean region (see Allen et al. 1999 and Chapter 4). These climatic oscillations (characterized by episodes of warm and cold, wet and dry conditions) had major effects on river behaviour— both indirectly, through changes in vegetation cover (one of the primary controls of river basin hydrology and sediment supply) and directly, through changing atmospheric circulation and associated weather systems that give rise to extreme hydrological events, such as floods and droughts (Greenbaum et al. 2006). 2. In the more tectonically active parts of the Mediterranean, such as south-east Spain and much of Greece and Turkey (see Vita-Finzi 1986, and Chapter 1), the Late Quaternary Period is of sufficiently long duration to allow observation of the impact of endogenic processes (including crustal movements and volcanic activity) on river network configuration and rates of channel incision—as well as on long-term patterns of fluvial

320

Mark Macklin and Jamie Woodward

Fig. 11.1. A building under excavation at the archaeological site of Olympia in the valley of the Alfíos River in the western Peleponnese, Greece. This is the ancient site of the Olympic Games and the entire site was buried by alluvium from the Kladeos River which is a small right bank tributary of the Alfíos. The doorway on the right and the room on the left are still filled with alluvium (photo: Jamie Woodward).

sedimentation and storage (Harvey and Wells 1987; Collier et al. 1995; Westaway et al. 2004). Tectonic activity is also an important control on the location and development of badlands as discussed in Chapter 1. 3. This is a key period for the development of our own species. Anatomically modern humans entered Europe for the first time via the eastern Mediterranean around 100,000 years ago and, by about 30 to 25 ka, Neanderthals had disappeared from the archaeological record. The long-term dispersal and seasonal subsistence patterns of these early humans were commonly focused on coastal environments and in river valleys. Thus, many important Middle and Upper Palaeolithic sites are found in fluvial settings (e.g. Higgs and Vita-Finzi 1966; Gamble 1986; Bailey 1997) and the reconstruction of Late Quaternary river environments is often key to their interpretation (Pope and van Andel 1984; Bailey et al. 1990; Macklin et al. 1997).

4. The Late Quaternary includes the present Holocene interglacial during which human activity has had an increasing impact on the Mediterranean environment (Chapter 9) including its river systems. Indeed, one of the most challenging and long-standing problems that confronts the study of Holocene river development in the region is how to isolate the effects of human activity on the nature and rate of fluvial erosion and deposition (Vita-Finzi 1969; Butzer 1980; Lewin et al. 1995; Bintliff 2002). Over the past decade or so, this has been a major research focus facilitated by the increasing availability of continuous land- and ocean-based proxy-climate records that document significant Holocene climatic variability in the Mediterranean region extending into the historical period (e.g. Casford et al. 2001; Sadori 2001; Roberts et al. 2001). 5. In order to identify and understand both the processes that govern river behaviour and causality

Rivers and Environmental Change

relationships in fluvial systems, it is essential that alluvial units are reliably dated. Increasingly robust fluvial geochronologies based on radiometric dating methods such as luminescence and uranium series are now available for a growing number of Mediterranean river basins and these techniques have led to new insights into Middle and Late Pleistocene river behaviour (Macklin et al. 2002; Woodward et al. 2008). In addition, the development of accelerator mass spectrometry has revolutionized the dating of terminal Pleistocene and Holocene age fluvial deposits by the virtue of being able to date very small samples of organic material that could not have been dated using conventional radiocarbon dating methods (Woodward et al. 2001; Thorndycraft and Benito 2006). This has produced a rapidly expanding database of radiometrically dated Late Glacial and Holocene fluvial units that, in conjunction with new high-resolution proxy climate records, have enabled the relationship between river behaviour and climate change to be documented and interrogated in increasing detail over both space and time (e.g. Abbott and Valestro 1995; Woodward et al. 2001; Macklin et al. 2002; Thorndycraft and Benito 2006; Zielhofer et al. 2004). This chapter is divided into five sections. The first is a review of Mediterranean river systems and related environmental change research that highlights some of the major themes that have emerged over the last five decades or so. The second section presents a new typology of Late Quaternary Mediterranean river development, with particular reference to the tectonic and geological settings of river basins and the effects of Pleistocene environmental change—including glaciation. In this section, Mediterranean river responses to environmental change are examined over the last glacial– interglacial cycle (c.130 to 10 ka) where the effects of both orbitally driven and millennial-scale climatic variability are evident. The third part of this chapter examines the Holocene record (the last c.11,500 years)—where human impact and climatic oscillations were the key drivers of change within river basins. Fourthly, we examine river behaviour during the last neoglacial cycle (sensu Rumsby and Macklin 1996), which started c.1, 000 years ago and includes the socalled ‘Medieval Warm Period’ and ‘Little Ice Age’ (LIA). The final section explores direct human modifications to river channel systems that have been especially prominent over the last 150 years or so and especially in the decades since the Second World War. Chapter 8 focuses on the present-day hydrology, river regimes, and sediment load of Mediterranean catchments and

321

includes a discussion of the implications of the Water Framework Directive for river systems in the European Mediterranean.

Models of Mediterranean Quaternary River Development Research conducted on Late Quaternary river histories in the Mediterranean up to the end of the 1960s was carefully reviewed and synthesized by Claudio Vita-Finzi in his acclaimed book, The Mediterranean Valleys: Geological Changes in Historical Times, published in 1969. Vita-Finzi’s study was particularly important because it was firmly grounded on almost a decade of fieldwork around the entire Mediterranean basin, extending from south-west Spain to the Jordan Valley and the Dead Sea. Using stone tools, pottery, coins, and other archaeological materials incorporated within alluvial sequences, for dating control, together with a very limited number of radiocarbon dates (five in total), he constructed a correlation scheme for Late Pleistocene (50 to 10 ka) and historical (post-Roman) alluvial units in the Mediterranean that he termed the ‘Older’ and ‘Younger’ Fills, respectively. Figure 11.2a shows a schematic crosssection of a valley in Epirus, north-west Greece where Vita-Finzi worked with the Cambridge archaeologist Eric Higgs in the 1960s (Higgs and Vita-Finzi 1966). This diagram developed from work on the Pleistocene and Holocene alluvial record in the Louros Valley near the Palaeolithic rockshelter site of Asprochaliko and it was instrumental in the development of Vita-Finzi’s Mediterranean-wide model published in 1969. It shows a prominent high terrace surface formed in a range of Pleistocene sediments—including strongly-weathered fan sediments or red beds—as well as trunk stream coarse-grained alluvial sediments that commonly interdigitated towards the valley sides with coarse limestone screes. Tufa deposits (Chapter 10) and cemented screes were common in this Older Fill. The historical or Younger Fill is set within the Pleistocene sediments in a narrow valley-floor setting created by Holocene incision and is dominated by fine-grained sediments that show little or no evidence of alteration by weathering. This phase of predominantly fluvial sedimentation has a much less variable sedimentology than the Older Fill and, according to the Vita-Finzi model, it was deposited after the Roman Period following a prolonged phase of early and mid-Holocene incision (Figure 11.2b). The thick deposits of fine-grained alluvium from the Kladeos River, that buried the ancient site of the Olympic Games at Olympia in the lower Alfios Valley of the Peloponnese,

322

Mark Macklin and Jamie Woodward metres

(a)

100

Cemented scree

Modern scree Pebble beds 50

Fan te

rrac e (R

ed B

rrace

River bed

eds) Tufa

Fan te

ds) ed Be

(R

Historical fill 0 Bedrock

Deposition

(b)

Aterian LevalloisoMousterian Oranian Capsian

Erosion

Medieval Roman

Acheulean

Neolithic

80

00

00

0

0

0

00

,0

,0

,0

40

20

00

00

00

0

0

0

0

00 ,0 10 0 00 8,

4,

2,

1,

80

40

20

10

Years ago

Fig. 11.2. (a) The Quaternary sediments of the Louros Valley in Epirus, north-west Greece (modified from Higgs and Vita-Finzi 1966). The high fan terrace is the surface of the Older Fill. These Pleistocene sediments are commonly strongly weathered and were also referred to as ‘Red Beds’. The historical fill shown in the diagram is the Younger Fill (see the main text for fuller descriptions). (b) The temporal record of channel and floodplain deposition and incision associated with the Older and Younger Fill model put forward by Vita-Finzi (1969).

are a classic example of Vita-Finzi’s Younger Fill (Figure 11.1). Vita-Finzi’s Older and Younger Fill model is probably best remembered for reducing alluvial and related colluvial deposits in Mediterranean river systems into two timed-confined eras of valley floor aggradation (see Bintliff 2002, for an insightful recent review of VitaFinzi’s contributions to Mediterranean geoarchaeology). It is clear, however, on close reading of The Mediterranean Valleys that he did in fact recognize, on the basis of the preservation of multiple alluvial terraces associated with

both the Younger (Vita-Finzi 1969: 91) and Older (ibid. 92) Fills, that each comprised several erosion and sedimentation cycles. The principal reason why Late Pleistocene and historical age alluvial units were conflated by Vita-Finzi, and by many subsequent workers, into single phases was because of the poor precision and accuracy of the dating control available at that time. Indeed this work preceded, by at least two decades, the development and application of appropriate geochronological tools (other than radiocarbon) for dating Late Quaternary alluvial

Rivers and Environmental Change

sequences. With hindsight, attributing both aggradation episodes (particularly the Younger Fill) solely to climate change seemed, even in the late 1960s, to stretch the empirical evidence, especially as high-resolution records of Late Pleistocene and Holocene climate change were not available at the time, and most reconstructions of human activity in the Mediterranean region were also not well constrained either spatially or temporally. VitaFinzi’s book was the first synthesis of Mediterranean alluvial geochronology and it set up a series of competing hypotheses of likely causative factors controlling longerterm river behaviour. This is arguably its key legacy as these bold ideas structured subsequent research and debate not only in the Mediterranean (Lewin et al. 1995; van Andel et al. 1990; Bintliff 2002) but also in the rest of Europe (Macklin et al. 2006a). There have been many advances in the four decades since Vita-Finzi’s model was first published. During the 1970s, and especially in the 1980s, a number of research programmes both in the Greek Islands (e.g. Renfrew and Wagstaff 1982) and in mainland Greece (Paepe et al. 1980; Pope and van Andel 1984; Bintliff and Snodgrass 1985), brought together interdisciplinary teams (including geomorphologists and soil scientists) coordinated by archaeologists to undertake regional Holocene landscape history projects that produced new and, in some cases, highly detailed data on human settlement in both the prehistoric and historical periods. Such information had been lacking in Vita Finzi’s and other earlier studies (e.g. Judson 1963; Raphael 1973), but these later investigations were themselves similarly limited as alluvial units were still not precisely dated and regional Holocene climate records against which supposed causative factors could be tested were also still not available (Bintliff 1992; Macklin 1995). From the mid-1980s, however, primarily as the result of the development and application of new dating techniques such as luminescence, uranium-series, and AMS radiocarbon dating, and through the reconstruction of continuous, high-resolution records of Late Pleistocene climate change (e.g. Allen et al. 1999), the chronology and controls of Pleistocene age alluvial units in the region were reassessed and in some cases radically revised (Pope and van Andel 1984; Bailey et al. 1990; Lewin et al. 1991; Fuller et al. 1998; Rose and Meng 1999; Kelly et al. 2000; Rowan et al. 2000; Woodward et al. 2001). These studies demonstrated there were multiple ‘Older Fill’ type Pleistocene alluvial units in most Mediterranean catchments and that Mediterranean rivers had been particularly responsive to short-term (millennial-side) climate change during

323

the Last Glacial period. More recently, there has been a growing interest in the incidence, cause, and geomorphic impact of large flood events in the Mediterranean region and the relationship between climate change and flood frequency and magnitude. Very large boulders (>2 m) on the channel bed of the Voidomatis River in the Vikos Gorge of north-west Greece are shown in Figure 11.3. It is clear that high magnitude flood flows with very high critical shear stresses are required to move such particles and direct observations of rare catastrophic flood flows in the Mediterranean (e.g. Piegay and Bravard 1997) have fostered an increased interest in their potential role in Pleistocene and Holocene sediment fluxes and valley floor development. Studies of the occurrence and impact of large floods in Mediterranean river catchments have been undertaken on Late Pleistocene (Woodward et al. 2001; Greenbaum et al. 2006), Holocene (Benito et al. 2003a ) and historical (Barriendos Vallve and Martin-Vide 1998) timescales and particularly detailed reconstructions are beginning to emerge for the Late Holocene (Greenbaum et al. 2000; Benito et al. 2003b) and historical (Maas and Macklin 2002) periods. Over the last decade a number of our own doctoral research students have made important contributions in this area with field investigations conducted in the Peloponnese (Christopoulos 1998; Zorzou 2004), Crete, (Maas 1998; Noble 2004), north-west Greece (Hamlin 2000), and Corsica (Hewitt 2002).

Towards a Late Quaternary Evolutionary Typology of Mediterranean River Basins The present-day morphology of river basins that drain to the Mediterranean Sea, as well as the nature and configuration of Late Quaternary fluvial sedimentary sequences found within them, results from the interplay between four major series of relief-forming factors (Macklin et al. 1995; Collier et al. 1995). These are crustal mobility (directed in both horizontal and vertical directions), rock type, periodic climate, and sea-level change and, in more recent times, human action (Chapter 1). The effects of short- and long-term climate change and anthropogenic activity will be discussed later in this chapter. In this section the roles of tectonics and structural controls on Late Quaternary river development are considered, together with the influence (directly and indirectly) of Pleistocene glaciation on fluvial regime, catchment sediment supply, and river sediment fluxes. Glaciation was particularly important in the mountain catchments of the northern part of the Mediterranean (Chapter 12; Hughes et al. 2006a). Taking into

324

Mark Macklin and Jamie Woodward

Fig. 11.3. Large boulders (>1.5 m) on the bed of the channel of the Voidomatis River in the Vikos Gorge of north-west Greece. Mediterranean mountain river systems commonly display good slope-channel coupling and the steep scree slopes in the background deliver coarse sediment directly to the channel system (photo: Jamie Woodward).

consideration tectonic-setting, in conjunction with the presence or absence of Late Pleistocene (or earlier) glaciation in a catchment, a simple evolutionary typology of Late Quaternary fluvial systems in the Mediterranean basin is outlined.

Tectonic Setting and River Basins The Mediterranean basin forms the boundary zone between the Eurasian, African, and Arabian Plates (Chapter 1). The interaction of these plates has produced the Alpine fold belt that extends across the Mediterranean from Gibraltar to the Middle East. The region has an extremely complex and variable structure that comprises a number of microplates that have, in some cases, very different tectonic and resulting stratigraphic histories to the adjoining Eurasian and African cratons (Dewey et al. 1973). The result of this long-term collisional history, as far as Late Quaternary river basin

development and fluvial environments are concerned, has been to produce, in general terms, three rather distinct tectonic settings around the basin for river systems that drain to the Mediterranean Sea (Macklin et al. 1995): 1. The first of these is the Precambrian African plate underlying much of North Africa. With the exception of the coastal areas of Cyrenaica in northeast Libya, this is mainly a low elevation desert environment with very infrequent seismic activity (Chapter 16). In the eastern Mediterranean, however, it is diversified by rifting in the Sinai and the Jordan Valley. 2. The second is the folded and partly metamorphosed Variscan massifs of the Iberian Peninsula, Corsica, and Sardinia in the western Mediterranean. In eastern Spain, flat-lying or gently folded Mesozoic and Cenozoic sediments cover these

Rivers and Environmental Change

massifs. These areas are seismically quiescent and rarely affected by earthquakes (Chapter 16). 3. The third and largest landscape element in the Mediterranean is the Alpine fold-and-thrust belt that runs across the entire region from the Maghreb and Pyrenees in the west, through the Apennines, Sicily, and Alps proper in the central Mediterranean and extending eastward to Greece and Turkey (Chapter 1). From the point of view of river development, an important characteristic of all of these areas is their high relief, the active nature of tectonics throughout the Quaternary, and the availability in many catchments of terrains formed in mechanically weak, readily erodible lithologies such as flysch and marl (Woodward 1995). Tectonics and bedrock lithology have exerted a significant influence over fluvial systems in the Mediterranean basin because of their influence upon large-scale drainage basin morphology (size and shape) and river long-profiles and sediment fluxes—as shown for mainland Greece by Collier et al. (1995). The development of Mediterranean river systems during the Pleistocene is presented below by tectonic setting (1–3). This is followed by a Mediterranean-wide discussion of the Holocene record of river basin system response to environmental change.

Tectonic Setting 1: The Precambrian Plate of North Africa Along the presently tectonically quiescent coastline of North Africa, river systems in northern Libya that drain to the Mediterranean Sea display Late Quaternary fluvial landforms and sedimentary sequences which are generally similar over a distance approaching 1,000 km—from Tripolitania in the west (Hey 1962; Vita-Finzi 1969; Anketell et al. 1995; Gilbertson et al. 1996) to Cyrenaica in the east (Vita-Finzi 1969; Rowan et al. 2000). This suggests that, in this region, fluvial geomorphic responses to environmental change over this period have been strongly conditioned by regional geology and Pleistocene base-level change histories. Two large-scale landform elements are present in this region: a coastal plain that ranges from about 5 to 150 km in width, which is flanked to the south by a limestone plateau that rises steeply from the plain in a series of fault-guided escarpments (Figure 11.4a) lying up to 500 m above sea level. The plateau is deeply dissected by ephemeral wadi systems that emerge from the uplands onto the coastal plain as a series of coalescing alluvial

325

fans. Due to progressive base-level fall since the Middle Pleistocene, these fans have a telescopically segmented form (sensu Blissenbach 1954) because the wadi channels have cut down into pre-existing fan sediments over several stages of development. The limestone uplands contain numerous inter-montane basins and these are commonly infilled with up to 30 m of Pleistocene alluvial and colluvial gravels. Schematic block diagrams of the Late Quaternary history of a typical Tripolitanian wadi produced by VitaFinzi (1969) (Figure 11.4b) show virtually identical sedimentary architectures and terrace morphologies to those recorded by Rowan et al. (2000) in Cyrenaica more than 500 km to the east. A detailed study of coastal alluvial fan development in the catchment of the Wadi Zewana, in north-eastern Libya, has been carried out in conjunction with an investigation of a series of alluvial fills within the main valley. This provided the first radiometrically dated geochronology of Middle and Late Pleistocene river activity in one of the least researched areas of the Mediterranean (Rowan et al. 2000). The Wadi Zewana drains a small (10 km2 ) catchment that meets the coast to the east of Tolmeita in Cyrenaica. The headwaters of this system lie above 350 m on the northern slopes of the Jebal Akhdar. Last interglacial bioclastic beach rock and overlying aeolianite (cemented dune sand) exposed in a coastal cliff eroding the toe of Zewana fan (Figure 11.4a) are found just above present sea level and indicate little or no uplift during the Late Pleistocene. Three major Late Pleistocene alluvial terraces have been mapped and radiometrically dated within the valley of the Wadi Zewana. The oldest and most extensive fill forms a prominent paired terrace approaching 25 m above the bed of the present river channel (Figure 11.5a). It comprises a sequence of generally flat-bedded fluvial sub-rounded gravels and sandy silts within which poorly sorted and angular colluvial gravels that thicken towards the edges of the valley are interposed (Figure 11.5b). A conspicuous feature of both fluvial and colluvial gravel units in this fill is a fine-grained matrix of terra rossa that has been eroded from the adjacent limestone slopes. This is a strongly weathered red soil found on limestone terrains across the Mediterranean (Chapter 6). Vita-Finzi (1969) noted a similarly high content of reworked terra rossa in Pleistocene wadi fills of Tripolitania. Uranium series, ESR, and OSL ages show that accelerated slope erosion, soil and regolith stripping, and valley floor infilling began sometime between c.80 and 70 ka and ended before c.42 ka when the river trenched back down to bedrock before refilling the valley floor. To establish the broader context of this record, these ages are shown

(i)

(ii)

(vii)

(iii)

(iv)

(v)

(vi)

Fig. 11.4. (a) View looking across the coastal plain in Cyrenaica (north-east Libya) showing the lower course of the Wadi Zewana in the foreground (photo: Mark Macklin). (b) Block diagrams showing the evolution of the Late Pleistocene and Holocene alluvial stratigraphy at a trunk stream tributary confluence in wadi systems in Libya (modified from Vita-Finzi 1969). The geomorphological and hydrological characteristics of these Libyan wadis are typical of many wadi systems that drain the coastal regions around the south-eastern Mediterranean, including the northern Sinai and the southern Levant.

Rivers and Environmental Change

327

(a)

(b)

Fig. 11.5. (a) The lower course of the Wadi Zewana showing a c.25 m thick exposure in Late Pleistocene alluvium. (b) A section showing the Late Pleistocene river sediments with coarse-grained angular gravels exposed at the base (photos: Mark Macklin).

on Figure 11.6 along with dated alluvial units from 16 other river basins across the Mediterranean and including key study areas discussed in this chapter. Significant slope erosion in the Zewana catchment occurred at c.23 ka and this was probably the precursor of, or perhaps coeval with, a phase of sedimentation on the coastal plain that had ended around 18 ka (Figure 11.6). The third major alluvial terrace within the Wadi Zewana rises to 12 m above the present channel bed, and a date of c.12.5 ka from the middle part of the unit, and a buried Roman cross-wadi wall c.3 m below the surface of the terrace, indicates sedimentation from the end of the Late Pleistocene until incision in the post-Roman period followed by limited valley floor refilling. Although Vita-Finzi did not have radiometric dates either for the Late Glacial or for the pre-Roman Holocene in north-west Libya, as with the earlier Late Pleistocene alluvial record, there are strong parallels between the evolution of wadi systems in Tripolitania and Cyrenaica in this period. The Wadi Zewana record suggests that Late Quaternary valley floor development in the tectonically less active parts of North Africa was characterized by an unprecedented phase of regolith and bedrock erosion that took place around 69 ka. This stripped last interglacial and earlier weathered material (terra rossa) from hillslopes to leave the bare limestone uplands of today. Although the precise duration of this phase of erosion has not yet been determined, it is clear that as the bedrock became progressively exposed in the Zewana catchment, this material was increasingly supplemented by mechanically weathered limestone gravel, resulting in aggradation along the entire trunk stream as well

328

Mark Macklin and Jamie Woodward Age (ka) 10

20

30

40

50

60

70

80

90

100

110

120

130

Guadalope, Spain, 41°00'N Guadalope, Spain, 40°50'N Bergantes, Spain, 40°45'N Bergantes, Spain, 40°40'N Voidomatis, Greece, 40°40'N Torrente d’es Coco, Mallorca, 39°40'N Loutro, Greece, 37°25'N Río Aguas, Spain, 37°10'N River Evrotas, Greece, 37°00'N Omalos, Crete, 35°19'N Samaria, Crete, 35°17'N

IRSL/OSL river terrace

Dating control and depositional environment

Rapanas, Crete, 35°15'N

IRSL/OSL fan U/Th river terrace U/Th fan

Aradena, Crete, 35°13'N

AMS 14C river terrace

Oued es Seffaia, Tunisia, 34°00'N

Incision

Alluviation

Wadi Zewana, Libya, 32°30'N Tissint feija, Morocco, 29°55'N Seyad feija, Morocco, 29°04'N

Total arboreal pollen (%)

100 80 60 40 20 0

d18O (‰)

-34 -36 -38 -40 -42 -44

0

10 1

20 2

30

40 3

50

60

70 4

80

90

100

110

5a

5b

5c

5d

120 5e

130 6

Oxygen Isotope Stage

Fig. 11.6. Dated alluvial units in river systems across the Mediterranean region between c.130 and 10 ka shown in association with two proxy climate records. The river basins and reaches are ordered by latitude from north to south. The legend shows the dating methods that have been employed in each case and the error bars for each date are also shown. The dated alluvial units come from papers cited in this chapter and the proxy climate records are from Tzedakis et al. (2002).

Rivers and Environmental Change

as in the larger tributaries (Figure 11.5). Following this major phase of aggradation, from c.42 ka the Wadi Zewana became progressively supply-limited with later phases of valley floor sedimentation resulting from the erosion and ‘cannibalization’ of the c.69 ka alluvial fill (see Lewin and Macklin 2003). The period shortly after c.69 ka along the southern Mediterranean littoral in Libya was a time of marked geomorphic change. Indeed, it represents probably the most significant landscape transformation that has occurred in this region for the last 100,000 years and perhaps longer (Figure 11.6). This period lies close to the Marine Isotope Stage (MIS) 5/4 transition which is dated to c.70 ka on the SPECMAP timescale and to about 75 to 80 ka in the GRIP ice core (Dansgaard et al. 1993 and Figure 11.6). It was marked by a sharp decline in sea-surface temperatures in the North Atlantic by as much as 5–6 o C (Bond et al. 1993) and the replacement of mixed woodland in southern Europe by open steppe (Tzedakis et al. 2002 and Chapter 4). This geomorphic ‘event’ set the boundary conditions for all subsequent phases of valley floor erosion and sedimentation in the region, including those that occurred during the Holocene and historical periods. It is clear that a major threshold was crossed at around 70 ka in terms of both the rates and types of hydrological and geomorphological processes operating within the river basins of the region. Before 70 ka geomorphic activity was relatively subdued in the catchment. After this date, however, there was a step-change increase in slope erosion and concomitant valley floor aggradation that was probably triggered by changes in vegetation type and cover—themselves related to altered precipitation regimes. Thereafter the Zewana basin became more responsive to a series of major climatic fluctuations at c.42, 23–18, and 12.5 ka that are recorded in marine, ice core, and pollen records (Figure 11.7).

Tectonic Setting II: The Hercynian Massifs and Foldbelts The Hercynian massifs and fold belts that form much of the central and eastern part of the Iberian Peninsula, together with the Hercynian massifs of Corsica and Sardinia that were involved in Alpine folding, comprise the second tectonic terrain with respect to a morpho-structural classification of Mediterranean river basins and their Late Quaternary development. Located away from the main plate boundaries, these areas have had relatively little seismic activity over the last 130,000 years (Chapter 16). However, gradual long-term regional uplift has created high basin

329

relief (34% of peninsular Spain lies between 800 and 2,000 m above sea level) and, with extensive outcrops of sedimentary rock susceptible to mechanical breakdown and erosion, these landscapes provide ideal conditions for the formation and preservation of river terraces (Macklin and Passmore 1995; Santisteban and Schulte 2007). Glacial moraines and Pleistocene river terraces are well preserved in the mountains of Corsica (Conchon 1986), but these records are not yet securely dated. One of the most studied and securely dated Quaternary river terrace sequences in the Mediterranean is that of the River Guadalope and its major tributary the River Bergantes in north-east Spain that drain the Iberian Cordillera and flow northwards into the Ebro basin (Macklin and Passmore 1995; Fuller et al. 1996, 1998; López-Avilés et al. 1998). High resolution geomorphological mapping in the lower part of the Bergantes Valley has identified five major river terraces at 20–25 m, 16 m, 10–12 m, 5–8 m, and 2–3 m above present river level. Luminescence dating of alluvial fills demonstrates trunk river aggradation at c.110 ka, c.35–40 ka, c.25– 7 ka, c.10–13 ka, c.7–8 ka, and c.3 ka, respectively. In broad terms, the scale of Late Pleistocene fluvial aggradation events varies considerably—with the oldest (c.110 ka) being the largest and with each subsequent event becoming progressively smaller. It is evident that differences in both the scale and pattern of Late Pleistocene fluvial sedimentation were controlled primarily by the degree of coupling between tributaries and the axial river and the amount of sediment supplied from local slopes and tributary catchments. The period around 110 ka saw up to 20 m of trunk river aggradation in the Bergantes Valley associated with exceptionally high rates of erosion and sediment delivery during the cold episode of Marine Isotope Substage 5d (Figure 11.6). This represents a ‘tipping point’ in catchment evolution and Late Quaternary river development in the region and appears to be directly analogous to the exceptional phase of drainage basin transformation at c.70 ka recorded in North African Mediterranean river systems discussed above. Both the c.110 ka and 70 ka phases of accelerated geomorphic activity coincided with forest decline and the development of steppe-dominated landscapes as shown by pollen records in southern Europe (e.g. Allen et al. 1999; Tzedakis et al. 2002) (Figure 11.6). Similar vegetation changes are evident in later stadial events, but the magnitude of the geomorphic response in north-east Spain during Marine Isotope Substage 5d suggests that climatic forcing of catchment erosion, sediment yields, and fluvial system dynamics were amplified in comparison to later cold

10

20

Age (ka) 40

30

50

60

Guadalope, Spain, 41°00'N Guadalope, Spain, 40°50'N Bergantes, Spain, 40°45'N Bergantes, Spain, 40°40'N Voidomatis, Greece, 40°40'N Torrente d’es Coco, Mallorca, 39°40'N Loutro, Greece, 37°25'N Río Aguas, Spain, 37°10'N River Evrotas, Greece, 37°00'N IRSL/OSL river terrace IRSL/OSL fan

Dating control and depositional environment

Omalos, Crete, 35°19'N

U/Th river terrace U/Th fan AMS 14C river terrace

Samaria, Crete, 35°17'N

Alluviation Incision

Rapanas, Crete, 35°15'N Aradena, Crete, 35°13'N Oued es Seffaia, Tunisia, 34°00'N Wadi Zewana, Libya, 32°30'N Tissint feija, Morocco, 29°55'N Seyad feija, Morocco, 29°04'N - 34 - 36

GRIP d18O (‰ SMOW)

- 38 - 40 - 42 -44 40

Arboreal pollen (%)

20 0 40

Steppe taxa (%)

20 0 0

IRD (lithic grains g-1)

50 100

H1

150

H2

e

d

10 1

20 2

H3

H4

H5

c

H6

b

30

40

a

50 3

60 4

Oxygen Isotope Stage

Fig. 11.7. Dated alluvial units in river systems across the Mediterranean region between c.65 and 10 ka. The shaded columns show Heinrich Events 1–6 (see Chapter 4). The river basins and reaches are ordered by latitude from north to south. The legend shows the dating methods that have been employed in each case and the error bars for each date are also shown. The dated alluvial units come from papers cited in this chapter and the proxy climate records are from Roucoux et al. (2005). The five palynological sections (a to e) from Roucoux et al. (2005) are also shown.

Rivers and Environmental Change

periods, perhaps through a combination of exceptional rates of runoff and sediment supply. Subsequent phases of valley floor refilling in northeast Spain are coeval with Heinrich event 4 (c.36–41 ka) and the Younger Dryas Stadial (Figure 11.7). Heinrich event 4 is believed to have been one of the most extreme climatic events of MIS 3, judged by the major flux of icerafted detritus in the North Atlantic from the Laurentide ice sheet (McManus et al. 1998) and the reduction in the thermohaline circulation (Tzedakis et al. 2002). In comparison to substage 5d, however, there was limited sediment supply from tributary streams during these periods and material that was delivered from tributary catchments was generally fine-grained in calibre. As elsewhere in the Mediterranean, while phases of large-scale Late Pleistocene river aggradation in the Bergantes Valley are correlated with Heinrich events and with stades of the Dansgaard-Oeschger cycles (Macklin et al. 2002), the record of river response to sub-orbital-scale climate change is more complex and spatially variable. In the tectonically quiescent Mediterranean regions of North Africa and the Iberian Peninsula, environmental changes were characterized primarily by reorganizations within the fluvial sediment system—the most notable of which were major fluctuations in tributary stream sediment delivery, and changes in alluvial storage within high-order valleys. These adjustments occurred in both areas over extended periods. In the drainage basins in north-east Spain, these began during the transition to the last cold stage (MIS 5d to 5a) and in north-east Libya were initiated during the MIS 5/4 transition (Figure 11.6). Over timescales of 104 years and longer, sediment yields have therefore been highly variable and data from the River Bergantes demonstrate very clearly that some of the larger (>1200 km2 ) Mediterranean river systems have been highly sensitive to changes in climate.

Tectonic Setting III: The Alpine Mountain Belt The largest and most geologically complex tectonic terrain in the Mediterranean is that resulting from Alpine orogenesis. It comprises the mountain chains of the Pyrenees, the Alps, the Balkans, and Turkey related to continental collision, and fold-and-thrust belts in south-east Spain, the Maghreb, and the Italian Peninsula related to back arc opening (Chapter 1). These landscapes include the most seismically active areas of the Mediterranean (Chapter 16). In many mountain regions within this tectonic province, particularly along the northern Mediterranean littoral of Italy, the

331

former Yugoslavia, Greece, and Turkey, climatic cooling during the Pleistocene provided favourable conditions for glacier development (Chapter 12). Glaciers had a significant impact on sediment supply and catchment hydrology, and these effects have been observed in the Pleistocene fluvial record in river valleys draining glaciated drainage basins in north-west Greece (Lewin et al. 1991; Woodward et al. 1992, 1995 2008), northwest Slovenia (Bavec et al. 2004) and north-east Spain (Sancho et al. 2003). It is therefore logical for the purposes of developing an evolutionary typology to subdivide Mediterranean river basins within the area affected by Alpine orogenesis into those catchments which supported important Pleistocene cirque and valley glaciers and those that did not. The Pleistocene glacial history of the Mediterranean mountains has recently been reviewed by Hughes et al. (2006a ) and is examined in detail in Chapter 12.

Glaciated River Basins in the Alpine Fold Belt The widespread occurrence of glacial landforms and sediments in the headwaters of many Mediterranean river basins has been known for well over a century (Cviji´c 1900; Niculescu 1915; Almagià 1918; Chapter 12). However, outside north-east Spain (Pyrenees; Sancho et al. 2003), north-west Slovenia (South Julian Alps; Bavec et al. 2004) and, most notably, the Voidomatis River basin in the Pindus Mountains of northwest Greece (Bailey et al. 1990; Lewin et al. 1991; Woodward et al. 1992, 1995 2001; Macklin et al. 1997, 2002; Hamlin 2000; Hamlin et al. 2000; Woodward et al. 2008), the downstream impact of this glacial activity upon Pleistocene river development is still relatively poorly known. A substantial body of geomorphological, sedimentological, and geochronological data has been assembled for the Middle and Late Pleistocene history of the Voidomatis River basin (384 km2 ). A large area of the basin’s headwaters lies in the high karst of the Pindus Mountains (Mount Tymphi, 2,470 m) and was subject to the build-up and decay of cirque and valley glaciers on at least three occasions over the last 500,000 years or so (Woodward et al. 2004; Hughes et al. 2006b). The glacial sequence on Mount Tymphi is now one of the best-dated in Europe and is discussed more fully in Chapter 12. In brief, the record comprises extensive Middle Pleistocene glacial sediments and landforms that have been dated to MIS 12 (between c.478 and 423 ka) and MIS 6 (between c.190 and 126 ka) using uraniumseries methods. In contrast, only small cirque and valley glaciers and rock glaciers developed during the last

332

Mark Macklin and Jamie Woodward

Pl e i s t o c e n e

Holocene

U8

Soil TL >150 ka

TL 28.2 ± 7 ka

Sand & silt

U6

Sample location

U5 U4

U/Th 113 ± 6 ka

U3 U/Th 80 ± 7 ka

U/Th 74 ± 6 ka

U/Th 53 ± 4 ka

U/Th 56 ± 5 ka

(a)

Gravel

TL 19.6 ± 6 ka

U7

U2

U/Th 24 ± 2 ka 25 ± 2 ka

U1

ESR 24.3 ± 2.6 ka 25.0 ± 0.5 ka 26.0 ± 1.9 ka

(b)

CaI AD 1420−1650

Present channel

(c)

Fig. 11.8. (a) The Pleistocene and Holocene fluvial stratigraphy in the middle and lower reaches of the Voidomatis River basin showing eight alluvial units including the modern channel complex. Based on a figure in Hamlin et al. (2000) using data from Lewin et al. 1991; Macklin et al. 1997; Hamlin et al. 2000). Key to dating methods: U/Th = uranium thorium, ESR = electron spin resonance, TL = thermoluminescence (both dates on the U6 and U3 soils are TL dates. U2 has been dated using radiocarbon). The photographs show (b) Unit 6 in the Lower Vikos Gorge at the location of soil profile A of Woodward et al. (1994) and (c) Unit 2 in the southern part of the Konitsa basin a few hundred metres downstream of Boila rockshelter (photos: Jamie Woodward).

cold stage or MIS 5d to 2 (between c.111 and 11.5 ka) (Hughes et al. 2006b). Late Pleistocene river sediments are well preserved in the lower reaches of the Voidomatis River and a detailed record of river aggradation and dissection, and its relation to headwater glaciation, has been reconstructed (Lewin et al. 1991; Macklin et al. 1997; Woodward et al. 2001, 2008). The fluvial record in the lower reaches of the Voidomatis River is shown schematically in Figure 11.8a along with the array

of dates that constrain the phases of fluvial sedimentation. Uniquely, the geochronology combines these four methods and sample types: uranium-series dating of calcite cements, thermoluminescence dating of fine-grained alluvial soils, electron spin resonance dating of fossil teeth, and AMS radiocarbon dating of charcoal. Alluvial units (U7 to U4) have a very similar lithological composition (Figure 11.8b)—they are dominated by coarse limestone gravels (>94%) and they contain a fine-grained matrix derived from the

Rivers and Environmental Change

glacial deposits in the catchment headwaters (Woodward et al. 1992; Hamlin et al. 2000). These four units represent the Aristi Unit sediments originally described by Lewin et al. (1991) that have been related to glacial activity in the catchment headwaters (Figure 11.8b). The Aristi Unit was later subdivided into four phases of aggradation and incision following detailed mapping of the lower reaches of the basin in the late 1990s and, crucially, through the application of uraniumseries dating to carbonate cements in the coarse gravel matrix (Hamlin et al. 2000) (Figure 11.8a). These ages are also plotted on Figures 11.6 and 11.7 along with other radiometrically dated alluvial units from across the Mediterranean. These phases of aggradation represent periods of enhanced sediment supply to the lower reaches from a range of sediment sources including: direct input from headwater glaciers, fluvial reworking of pre-existing glacial and alluvial sediments, and cold climate weathering of hillslopes. It is interesting to note that the Voidomatis River terrace sequence provides a more detailed archive of Late Pleistocene environmental change than does the glacial record in the catchment headwaters (Chapter 12). Unit U8 is the Kipi Unit of Lewin et al. (1991) and it has a very different lithological composition to the later, limestone-rich units. It is dominated by flysch and ophiolite gravels (Figure 11.8b) and has a flysch-rich matrix. It must therefore predate the first phase of glaciation in the catchment (MIS 12) and the TL date of >150 ka (Figure 11.8a) is a minimum age. U3 is the youngest Pleistocene alluvial unit and is equivalent to the Vikos Unit of Lewin et al. (1991). The lithology of this unit shows that it was deposited after the final period of Late Pleistocene glacial activity on Mount Tymphi and it overlaps with a phase of Late Glacial slackwater sedimentation recorded in the deposits of Boila rockshelter in the lower reaches of the Voidomatis River (Woodward et al. 2001) and shown in Figure 11.9. The postglacial record of river behaviour is dominated by progressive incision that is interrupted in the late Holocene by the development of a distinctive low terrace (U2) formed in fine-grained overbank sediments (Figure 11.8c).

Non-glaciated River Basins in the Alpine Fold Belt The Sorbas basin, part of the Betic Cordillera of southeast Spain, lies within a rapidly uplifting area, with Pliocene to recent average uplift rates of c.80 m Ma−1 (Mather 2000; Chapter 1). This basin was not glaciated during the Pleistocene and provides one of the bestdocumented examples of the effects of tectonics on Late Pleistocene and earlier drainage network development in the Mediterranean (Harvey and Wells 1987).

333

Fig. 11.9. The sediments exposed during the excavations at Boila rockshelter in the lower reaches of the Voidomatis River in northwest Greece (after Woodward et al. 2001). The bottom of the section shows the fine-grained slackwater sediments produced by large floods prior to the Late Upper Palaeolithic use of the site. The sediments above contain angular limestone clasts derived from the walls and ceiling of the rockshelter. The scale is in cm (photo: Jamie Woodward).

Accelerated headward erosion of the Río Aguas, stimulated by regional differential uplift, resulted in progressive reorganization of the drainage network by river capture. The most dramatic modification of the network was a major capture of the south-flowing proto Aguas/Feos by the east-flowing lower Aguas, an aggressive subsequent stream that was developing by headward erosion along the outcrop of a weak marl unit. Three pre-capture river terraces can be traced following the proto-drainage across the Sorbas basin and through the southern mountain ranges (Harvey and Wells 1987). Uranium series dating of pedogenic calcretes developed on the surface of these river terraces

334

Mark Macklin and Jamie Woodward

Fig. 11.10. The deeply incised valley floor and Quaternary terraces of the Río Aguas in south-east Spain (see Chapter 1). This reach is a few kilometres downstream of Sorbas and shows the Urra field centre on the high terrace (photo: Jamie Woodward).

has recently been undertaken (Candy et al. 2005). These workers argue that the terraces date to c.304 ka, 207 ka, and 70 ka, respectively, with capture of the Río Feos by the lower Aguas taking place some time between 69.8 and 67.9 ka. These dates, along with three further OSL dates published by Schulte et al. (2007) are also plotted on Figure 11.6. The dates for the two oldest terraces indicate that river incision of 10–15 m began towards the end of the interglacials of MIS 7 and 9, followed by limited refilling (c.5 m) of the valley floor. In the case of the c.70 ka river terrace, calcite-cemented root mats in the middle part of the gravel-dominated unit show that it comprises two aggradational phases; one dated to sometime before 77 ka and the second shortly after 77 ka (Candy et al. 2005) (Figure 11.6). Post-capture terraces (<70 ka) are clearly expressed in the Sorbas basin (Figure 11.10), but patterns of river incision and aggradation, as well as badland development and landslide activity, have been governed by their location relative to the river capture site (Harvey et al. 1995). Upstream of the capture, a 90-m drop in base level included headward incision that over the last 70 ka has migrated 20 km upstream. Largescale drainage network incision in the upper reaches

of the Aguas following the c.70 ka capture event has resulted in the development of steep and unstable slopes. These slopes have been especially prone to badland erosion (particularly in weaker lithologies) and largescale rotational landslides (Griffiths et al. 2002) that have resulted in local damming, ponding, and finegrained sedimentation in the Aguas system (Mather et al. 1991; Kelly et al. 2000; Chapter 1). Downstream of the capture site, along the lower Aguas, rates of incision during the Late Pleistocene have not been as high but there has been increasing valley widening through lateral river migration (Harvey et al. 1995) (Figure 11.10). The wider significance of Harvey and his colleagues’ work in the Sorbas basin for understanding river development in the tectonically active parts of the Mediterranean affected by Alpine orogenesis lies in its identification of the differential roles that climate, tectonics, lithological variability, and river capture have had in regional landscape development. Two different styles of Pleistocene river evolution can be recognized—each reflecting differing relationships between these factors (Harvey et al. 1995). The first style exhibits progressive incision, punctuated by episodic

Rivers and Environmental Change

aggradation and dissection in response to long-term uplift of the basin, and river terraces are weakly divergent in headwater areas becoming parallel to one other further down the system. The second dissection style reflects major base-level change, the local impact of which is governed and intensified by river capture. River terraces diverge to the point of capture and then converge downstream. Analysis of soil development on both pre- and post-capture fluvial units does not identify any time-progressive behaviour of individual terraces (Harvey et al. 1995). This is an important observation as it implies that switches between fluvial dissection and aggradation took place relatively rapidly and that these occurred throughout the system—almost irrespective of local tectonic context—in response to climatically controlled changes in sediment supply. In this case, the primary role of tectonics has been to control the context within which short-term river aggradation and incision processes have operated during the Late Quaternary. In those parts of the Mediterranean that have seen active volcanism during the Late Quaternary (Chapter 15), basalt flows have interacted with long-term river activity in a very distinctive way. For example, part of the Gediz River basin in western Turkey lies within the Kula Quaternary volcanic field and basalt flows have filled gorge reaches to depths of many tens of metres on several occasions during the Middle and Late Pleistocene. These events created major perturbations to the river’s long profile and forced the trunk stream to re-incise the valley floor (Westaway et al. 2003). In some reaches these basalt flows cap Quaternary fluvial sediments and, because the basalts can be dated using K-Ar and Ar-Ar methods, they can be utilized to generate age models for rates of regional uplift and long-term river behaviour. Westaway et al. (2003) have obtained a series of dates from basalt-capped river terraces to develop a model of Pliocene to Quaternary surface uplift for western Turkey. Using a range of methods, including geophysical survey, Branca and Ferrara (2001) have examined the interactions between drainage system development and volcanic activity on the slopes of Mount Etna in Sicily during the course of the Middle and Late Pleistocene. Between 170 and 100 ka, the main channel of the Alcantara River was periodically invaded by lava flows from peripheral activity of the central volcano. A major modification of the regional drainage network began around 45–40 ka as the main eruptive centre of the volcano (the Ellittico volcano) increased in size. Over the next 25–30 ka, the gradual expansion of this volcano led to a progressive northward diversion of the Alcantara

335

channel and to the capture of an adjacent river catchment (Branca and Ferrara 2001).

River Systems and Holocene Environmental Change The most contentious area of environmental change and river system research in the Mediterranean has centered on the interpretation of the Holocene record. VitaFinzi (1969) identified a single phase of alluviation in the Holocene that took place between the end of the Roman Period and c.400 BP as shown in Figure 11.2b. In the river basins that he observed, his Younger Fill consisted primarily of fine-grained sediments commonly underlain by basal gravels. He argued that this period of fluvial deposition was a result of increased erosion and flooding produced by a cooler and wetter climate across the Mediterranean during the Medieval Period (Figure 11.2b). Vita-Finzi’s (1969) interpretation of the Holocene alluvial record in the Mediterranean initiated a major debate that continues today. With hindsight, several weaknesses in Vita-Finzi’s thesis for Holocene river development can be identified. The dating control for the Younger Fill was based on a very small number of radiocarbon samples and it therefore proved impossible to precisely determine the onset of fluvial aggradation and subsequent incision in the way that he envisaged across the entire region. Indeed, many subsequent studies have shown much more complex Holocene records that include major phases of fluvial sedimentation and incision that took place much earlier than the single phase of post-classical alluviation represented by the Younger Fill. For example, Figure 11.11 shows a series of river terraces in the Torcicoda River valley of central Sicily where a major phase of fluvial sedimentation began in the Neolithic period before c.5.5 ka. In very general terms, it can be argued that VitaFinzi’s (1969) proposed mechanism of river response to Holocene climate change—involving large-scale shifts in atmospheric circulation across the Mediterranean— has been shown to be broadly correct by later studies. However, in the late 1960s there were no independent, high-resolution climate records to test his proposal for the timing and mechanism of Holocene river alluviation and erosion and this was still the case when VitaFinzi (1976) explored the potential significance of the latitudinal pattern of river basins and a diachronous response to Late Holocene climate change. Again, the very limited dating control for the fluvial records did not allow these novel ideas to be properly tested.

336

Mark Macklin and Jamie Woodward

Fig. 11.11. The Holocene alluvial sediments and terraces in the middle reaches of the Torcicoda River in central Sicily (photo: Jamie Woodward).

Finally, Vita-Finzi’s (1969) model did not incorporate the potential impacts of human activity in river basins during the course of the Holocene and this omission attracted widespread criticism. Other workers have argued that forest clearance and a range of farming practices—including ploughing and hillslope terracing—had very significant impacts upon river basin hydrology and the transfer of sediment from hillslopes to river channels (Chapters 6, 8, and 9). Brückner (1986) provides a good summary of the view that emerged in the two decades following the publication of The Mediterranean Valleys. He provides a long list of human impacts on the Mediterranean landscape that took place during the course of the Holocene including the expansion of settlements, the destruction of the protective vegetation cover, the cultivation of grain and olive trees, widespread goat-keeping, and overpopulation. Brückner (ibid. 16) argued that these human impacts ‘all contributed to the areal denudation and erosion which caused the filling up of the valley bottoms and the growth of deltas. This clearly demonstrates man’s

predominant part in the evolution of the physical environment in historical times. Compared to that, the influence of climate and tectonics were subordinate.’ In a review of The Mediterranean Valleys, Karl Butzer argued that the Older and Younger Fill model was an oversimplification and that alluviation episodes were not ‘universally contemporary’ across the region. He also argued that ‘the field evidence is seldom adequate to support the bold generalizations and hypotheses put forward’ (Butzer 1969: 53). Over a decade later, in a review of dating and correlation issues surrounding Holocene alluvial sequences in various parts of the Old World, Butzer (1980: 138) argued that, in Mediterranean river basins, ‘the dominant pattern is one of accelerated soil erosion in response to human misuse of the land— beginning locally during the late Bronze Age and becoming more universal after the economic decline of the Roman Empire’. In contrast to many publications that appeared around this time, Butzer (1980) also acknowledged the important role of climate in driving river basin processes—but he still argued a strong case for

Rivers and Environmental Change Olive trees

337

(a) Prosperity

Grain fields

(b) Depression

Maquis

Stream deposits

Fig. 11.12. The prosperity and depression model of slope stability and soil erosion modified from van Andel and Runnels (1987). See text for discussion. The photographs show (top) well-maintained stonewalled agricultural terraces and (bottom) a breached terrace wall with sediment movement downslope (photos: Jamie Woodward; both from the Ladonas catchment in the western Peleponnese, Greece).

the role of human action: ‘Climatic impulses may have favoured such geomorphic activation, but the variable timing, and the unprecedented extent and scope of these slope and valley changes within the Holocene record, remain inconceivable without a pre-eminently cultural impetus.’ During the 1970s and 1980s, several large-scale archaeological surveys were conducted in the Mediterranean and this period saw the emergence of modern landscape archaeology in the region and new perspectives on landscape change. Building on Butzer’s arguments, work in Greece by Tjeerd van Andel and others on the Argolid Peninsula of the eastern Peloponnese led to the development of a model of hillslope and valley floor dynamics driven by economic factors (van Andel et al. 1986; van Andel and Runnels 1987). This model proposed that socio-economic and demographic factors were the primary controls on hillslope processes and river alluviation and erosion. Figure 11.12 illustrates

their model of alternating prosperity and depression and their respective impacts on hillslopes and river channels. In short, this work argued that agricultural terraces were well maintained during periods of economic prosperity; with only limited sediment transfer from hillslopes to river channels. Phases of channel alluviation were associated with periods of economic decline when terrace walls became degraded and this promoted the transfer of runoff and sediment to valley bottoms (Figure 11.12). ‘In the absence of independent evidence for increased precipitation and runoff and for other, circumstantial reasons, we have attributed those [streamflood deposits] to neglect of a well-established system of terraces and gully check dams, with the inevitable vertical erosion, soil removal, and increased flooding of streams that such neglect entails’ (van Andel et al. 1986: 125). Despite the relatively poor dating control for many Holocene fluvial records, van Andel et al. (1990)

338

AD 2000

Mark Macklin and Jamie Woodward Larissa Euboea S. Argolid Olympia S. Crete Naxos Macedonia Nemea Argive Plain Elis Melos Peloponnese

?

?

0

? ? 2000

? ? ?

4000 Modern Agro-economic After initial farming 6000 BC

compared the data from several river basins across Greece (Figure 11.13) and argued that the observed pattern of river aggradation reflected the variable timing and impact of catchment disturbance that followed the introduction of different agricultural technologies. This analysis showed clearly that there were multiple phases of Holocene alluviation in Greece, but a rigorous evaluation of the causative factors was still not possible given the lack of independent and high-resolution proxy climate records. A large volume of research has been conducted in the lower reaches of Mediterranean river basins to examine the mechanisms, timing, and underlying controls behind coastal progradation in the postglacial period (Kraft et al. 1975). Marine embayments were transformed into floodplains during the Holocene; and reconstructions of coastal palaeogeography have proved crucial for the interpretation of archaeological sites such as ancient Troy and Ephesus (Kraft et al. 1980, 2007). Ancient Troy was located on an embayment close to the sea—whereas today it lies 6 km from

Fig. 11.13. Patterns of Holocene alluviation in Greece for the last 8,000 years modified from van Andel et al. (1990).

the Aegean coast (Kraft et al. 1980). The interactions between river deltas, coastal plains, and sea-level change in the Holocene are discussed more fully in Chapter 13. More recent research in the Mediterranean has focused on the significance of large flood events in shaping the Holocene fluvial record and has incorporated very detailed work on alluvial sequences within high resolution dating frameworks based on large numbers of radiocarbon dates. Benito (2003) has compiled a Holocene record of river aggradation and incision along with flooding episodes for parts of both the African and European Mediterranean (Figure 11.14). Region-wide river aggradation can be recognized in the Medieval Period, the Iron and Bronze Ages, the Neolithic, and Late Mesolithic. A high incidence of major floods is associated with both valley floor sedimentation and incision— and indicates the potential variability of river response to hydrological change in different geomorphological settings. Although this was an important review, the lack of dating precision for key areas still largely precluded

Cal.

C -14

1

1

2

2

+

Modern SubAtlantic

Medieval

2.8 kyr

4

4

SubBoreal

Bronze Age Calcolithic

5

+

–

+

General discharge (Q) characteristics

High variance Roman

3

–

Q

High Q Low Q

Iron Age 3

Q

Floods

–

Greece Aggrad.

+

Q

Floods

–

Italy Aggrad.

+

Q

Floods

–

France Aggrad.

Q

Floods

Archaeological sequence

Spain Aggrad.

Pollen chronozones

Floods

Age kyr BP

Aggrad.

Tunisia

Low Q High Q Low Q

3.5 kyr

High Q ?

5 kyr

High variance

5 6 7

6

5.7 kyr

Atlantic Neolithic

8 8

Boreal

9 PreBoreal 10 Y. Dryas

Epipalaeolithic/ Mesolithic

11 12

Low Q High Q Low Q Increasing Q Low Q

Fig. 11.14. Patterns of fluvial aggradation and flooding in five Mediterranean countries for the Holocene period (modified from Benito 2003). The solid bars show periods of aggradation and flooding and the arrows mark the onset of phases of channel incision.

Rivers and Environmental Change

10

High Q 7.2 kyr

7

9

Low Q

339

340

Mark Macklin and Jamie Woodward

secure process-based correlations with periods of rapid climate change and human impact. Following 14 C-based fluvial database development and meta-analysis techniques pioneered by Macklin and Lewin in the UK (Macklin and Lewin 2003; Macklin et al. 2005), this methodology is now beginning to be applied to Mediterranean river basins. For example, Thorndycroft and Benito (2006) have compiled and analysed radiocarbon-dated Holocene fluvial units in Spain with a particular emphasis on slackwater flood deposits (but also including dates from overbank fine sediments and coarse gravel facies) formed over the last 3,000 years (Figure 11.15a, b). Their recent analysis shows that fluvial activity in Spain over the last 11,000 years or so (Figure 11.15a) has been markedly periodic, but with the majority of fluvial units dated to the last 1000 years (Figure 11.15b). This probability-based assessment of the sedimentary flood record has allowed for more meaningful and robust correlations with documented historical floods (Figure 11.15c) and with high-resolution climate records from the North Atlantic (Figure 11.15d). Using this approach, Spanish Holocene flood deposits can be compared for the first time to those elsewhere in Europe and with lake level records and climate proxies from the North Atlantic (Figure 11.16). This comparison shows that six of the major flood episodes in Spain (identified with an asterisk on the far right of Figure 11.16)—particularly over the last 5,000 years—coincide with flooding in the northern part of Europe, and with periods of cooler climate and high lake level. These periods of cooler climate affect river basin processes throughout Europe, including the northern part of the Mediterranean basin (Macklin et al. 2006a ). Much of the work that followed in Vita-Finzi’s footsteps in the 1970s and 1980s was hindered by three major problems. First, despite the rhetoric, there were, in effect, only very limited improvements in dating control for the fluvial record and, in most cases, this prevented effective testing of the competing theories. Second, the long-term record of land use change remained poorly documented and rather poorly dated in most regions. The age of stone-wall terracing, for example—which represents a key modification of the hillslope sediment system across many parts of the Mediterranean—is very difficult to establish (Moody and Grove 1990; van Andel et al. 1990). Third, detailed proxy records of Holocene climate change have only recently become available for the Mediterranean region and it can be argued that a truly convincing Mediterranean-wide assault on the nature and controls of river response to Holocene envi-

ronmental change has not yet taken place. Wider application of the radiocarbon database approach across the Mediterranean region will allow the Holocene fluvial record to be evaluated against the high-resolution climate records and the existing archaeological and historical records of human activity in river catchments in a much more rigorous manner.

Mediterranean River Systems and ‘The Little Ice Age’ Climatic conditions in the Mediterranean region during the ‘Little Ice Age’ (LIA, c.AD 1450–1850) were generally cooler and wetter than present (Grove 1988; Grove and Conterio 1995; Grove 2001). There is a growing body of data to suggest that this was a very significant period of fluvial activity across the region with enhanced sediment supply from hillslopes to river channels and a greater frequency of large floods (Barriendos Vallve and Martin-Vide 1998). There is also evidence for glacier expansion in the headwaters of some Mediterranean river basins during this period (Hughes et al. 2006a; Chapter 12). A number of studies have successfully applied lichenometry to date coarse-grained flood sediments in steep-gradient mountain catchments. This has allowed the development of high-resolution flood histories for the last few centuries. This approach has provided new insights into river behaviour during the LIA as instrumental hydrological records are commonly too brief or absent. One such study has focused on the Aradena Gorge in south-west Crete. This steep fluvial system drains the southern slopes of the Lefka Ori Mountains and Maas and Macklin (2002) have used geomorphological evidence to compile a record of recent flood events in the gorge using lichenometry to date the coarse flood deposits. They have identified twelve periods within the last 150 years when floods were more frequent than the last 10 years or so of the twentieth century. (Figure 11.17a). The periods of enhanced flooding appear to coincide with negative or declining phases of the North Atlantic Oscillation (Chapter 3). Using boulder size as a proxy for flood magnitude, Figure 11.17b shows the seventy-five flood units plotted against estimated age. This analysis shows that the largest floods in this record took place in the middle of the nineteenth century towards the end of the LIA. The full record from the Aradena Gorge shows a decrease in flood magnitude in the second half of the nineteenth century, an increase in the size of floods up to about 1930, and then a progressive decline in the second half of the twentieth century. It is important to appreciate

0.08

Probability per year

(a)

All dates (n=74)

0.06

0.04

0.02

0 10000

8000

6000

4000

2000

0

1000

500

0

Calibrated years BP 0.05

(b)

3000 – 0 BP (n=49)

Probability per year

0.04

Alluvial: overbank facies

0.03

Slackwater flood deposits 0.02

Alluvial: gravel facies

0.01

0 3000

2500

2000

1500 Calibrated years BP

(c)

Low flood frequency

Historical floods in Spain (Benito et al. 1996)

High flood frequency

Two-sigma age ranges from individual SFD sedimentary units Overbank facies (mid-point dates)

0.02

Overbank facies (change dates) 0.01

0 3000

SFDs (bracketed phases of increased flood frequency) 2500

2000 1500 1000 Calibrated years BP

Probability per year

0

LIA

MWP

0.03

2850BP event

(d)

500

15

0.02

10

0.01

5

0 3000

Drift ice lithic indicators (%)

Probability per year

0.03

0 2000

1000

0

Calibrated years BP

Fig. 11.15. (a) A summed probability plot for radiocarbon dates from Holocene alluvial records in Spain (11 ka to present). (b) The boxed area in (a) showing probability plots for each of the three alluvial depositional contexts shown. (c) Summed probability plots based on the change and mid-point alluvial data sets and the bracketed slackwater flood deposits. (d) A summed probability plot for the radiocarbon dates from the slackwater sediments shown in relation to the North Atlantic drift ice index. LIA = Little Ice Age and MWP = Medieval Warm Period. (All figures modified from Thorndycraft and Benito 2006).

Ice rafting debris in the North Atlantic Ocean

British 14C dated flooding episodes

Dry phases Phases of higher in Lake Siles lake level in (southern Spain) mid-Europe

Spanish 14C dated flooding episodes

Polish 14C dated flooding episodes

0

570* 660* 860–880* 1040 1290–1310 1830–1950*

2,000

2550–2560 2730–2810* 3510–3540 4,000

Age (cal. BP)

4520 4840–4860* 5540–5640 5920 6250

6,000

6790–6820 7590 8,000 8400

9440–9530 10,000 10490 11160–11200 12,000 -5

0

5

4

2

0

-1 10

5

0

-2

2

1

0

-1

Probability difference (a.u.)

Fig. 11.16. Probability difference curves of radiocarbon dates associated with major flooding episodes in Great Britain, Spain, and Poland plotted next to the Ice Rafted Debris (IRD) record for the North Atlantic (after Bond et al. 2001) and hydrological records from lakes in southern Spain (Carrión 2002) and central Europe (Magny et al. 2003). Episodes of flooding in Europe are listed on the right-hand side and those episodes marked with an asterisk took place in all three regions (modified from Macklin et al. 2006).

Rivers and Environmental Change

(a) A

B

C

D

E

F

G H I

J

K

10

8

1

NAO index

6

4 0

2

–1 1840

1860

1880

1900

1920

1940

1960

1980

0 2000

1940

1960

1980

2000

Number of flood deposits per year

2

343

Year

1,500

(b)

B-axis (mm)

1,000

500

0 1840

1860

1880

1900

1920

Flood deposit age (year AD) Fig. 11.17. The flood record for the Aradena Gorge in south-west Crete between 1840 and 2000 based on lichen dating of coarsegrained flood deposits. (a) The number of flood deposits per year compared with a five-year running mean of the winter (December to March) NAO index (1840–1996). The grey columns show the eleven age classes (A–K) identified from the lichen age data (see text for further discussion). (b) The mean size of the ten largest boulders from each dated flood unit. These data can be used as a proxy for flood magnitude with the largest boulders transported during the biggest floods. Boulder size is shown as the mean value and range of the ten largest boulders (modified from Maas and Macklin 2002).

344

Mark Macklin and Jamie Woodward

that boulder size may also be a function of sediment supply from hillslopes, and the decline in boulder size may partly reflect less effective slope channel coupling (Maas and Macklin 2002). Interestingly, the gauged record of daily rainfall from 1955 to 1987 also indicates a greater frequency of sustained high-intensity storms during negative phases of the winter NAO. This work shows a clear relationship between climate change, flood frequency and magnitude, and sediment supply for the past 150 years and this pattern has been recognized in several studies of coarse-grained flood deposits in river channel systems across the Mediterranean region (Hamlin et al. 2000; López-Aviléz 1999; Hewitt 2002; Noble 2004). The historical flood record of the Rhône River provides valuable insights into the impact of the LIA on much larger-scale catchment dynamics. The Rhône drains a very large catchment (97,800 km2 ) that extends to the north of the Mediterranean zone of southern France. It provides a connection between the mountain catchments in the Alps—with their glacial and snowmelt regimes—and the Mediterranean lowlands and coastal zone of southern France where the floodplains and coastal zone have received large volumes of water and sediment during the course of the Pleistocene and Holocene. Glaciers expanded significantly in the Alps during the LIA (Grove 1988) and mass balance fluctuations impacted significantly on the Mediterranean region of the lower Rhône basin. From a detailed analysis of the historical flood record at the Arles gauging station between 1500 and 1995, Pichard (1995) has shown that the frequency of large floods decreased significantly after about 1860. Large floods were defined as those that reached a stage >5.25 m at Arles and this equates to a discharge of >7,000 m3 s−1 . These big floods were more frequent in the decades from 1651 to 1720 and 1751 to 1860. This flood-dominated regime was interrupted by periods of low flood frequency from 1721 to 1750 and from 1861 to 1995. The frequency of these large floods has fallen since the end of the LIA with eight to nine per decade between 1850 and 1900, four to five per decade between 1900 and 1950 and only two to three per decade between 1950 and 2000 (Pichard 1995; Arnoud-Fassetta 2003). Further south in the western Mediterranean, Hewitt (2002) has developed a flood history for the Figarella River (132 km2 ) that drains to Calvi Bay from the mountains of north-west Corsica. Here, flood deposits—including boulder splays, boulder bars, and boulder berms—are well preserved across four terrace surfaces and these have also been dated using lichenometry. Hewitt (2002) identified 18 flood units

that represent ten periods of enhanced flooding over the last four centuries, with the oldest units dated to between 1572 and 1595. Most of the flood units preserved in the Figarella record were deposited by floods that took place in the nineteenth century. In common with the record of large floods from the Rhône (Pichard 1995), the Figarella River has a much lower frequency of large floods in the twentieth in comparison to the nineteenth century. Hewitt (2002) has shown that the Figarella flood history shows a good agreement with the regional-scale record of large floods compiled from a range of data sources by Barriendos Vallve and Martin-Vide (1998) for ten river catchments that drain to the coast of north-east Spain (Figure 11.18). This figure compares flood records from documentary sources with geomorphological records from steep gradient river systems located in Corisca, Crete, and north-west Greece. Broadly speaking, five periods of enhanced flooding can be identified within these records: 1. The late sixteenth century (north-east Spain, southern France, and Corsica). 2. The late seventeenth to early eighteenth century (southern France). 3. Late eighteenth century (across the Mediterranean, but not in Crete). 4. The mid and late nineteenth century (across the Mediterranean). 5. The first half of the twentieth century (Corsica, north-west Greece, and Crete). The record from north-west Greece is the only one in this data set to show enhanced flooding in the second half of the twentieth century. It is important to appreciate that the records from France and north-east Spain shown in Figure 11.18 come from documentary and instrumental sources, whilst those from Corsica, north-west Greece, and Crete are from geomorphological evidence dated by lichenometry and these three may partly reflect the influence of preservation factors. The Corsican record goes back as far as the longest historical records so far published for the western Mediterranean.

Recent Human Modifications to River Channel Systems Over the last one hundred and fifty years or so, and especially since the Second World War, river channel systems in the Mediterranean region have been subjected to a range of modifications resulting from engineering works and resource exploitation. Impacts include

Rivers and Environmental Change

345

Fig. 11.18. Flood histories from five parts of the Mediterranean since AD 1500 based on geomorphological (solid shading) and documentary (hatched shading) records.

dam construction and reservoir development for hydroelectric power generation and flood control, as well as water resource development for irrigation and public water supply (Palanques et al. 1990; Surian 1999; Arnaud-Fassetta 2003). Reservoir sedimentation is now a major problem across the region (Lahlou 1988; Woodward 1995; Woodward and Foster 1997; Chapter 8) and this has led to river channel incision in downstream reaches (Surian and Rinaldi 2003). In the middle and lower reaches of many river catchments, engineers have drained wetlands, built flood embankments, straightened channels, and developed extensive water abstraction and irrigation schemes. The second half of the twentieth century has seen an increasing demand for sands and gravels for a range of civil engineering projects including road building and urban development. This has encouraged intensive gravel extraction from floodplains and from the active channel zone in many Mediterranean rivers (Nicholas et al. 1999). The response of several Italian rivers to river engineering projects and gravel extraction has been reviewed by Surian and Rinaldi (2003). They have shown that 3–4 m of channel incision is a common response in reaches of the Po, Arno, and Piave rivers and channel narrowing of more than 50 per cent has been observed. These channel responses are more rapid immediately following the disturbances and the larger rivers tend to have longer recovery times. Surian and Rinaldi (2003) propose a model for the main styles of

river channel response to recent human intervention in Italy with braided rivers adjusting through narrowing of the active channel zone with varying rates of incision, whereas single-thread river channels adjust via more pronounced vertical incision with variable degrees of channel narrowing. Precious and base metal mining in the Mediterranean has a very long history with mining activity in the Río Tinto of south-west Spain being amongst the oldest in Europe dating back more than 5,000 years. Historical mining activity has resulted in significant and widespread contamination in many Mediterranean river basins with floodplain sediment in mining-affected catchments presently acting as a major secondary source of pollution (Macklin et al. 2006b). Environmental problems also arise from present-day mining activities with the April 1998 Aznalcóllar tailings dam failure being one of the most serious river pollution incidents in Europe since the Second World War (HudsonEdwards et al. 2003; Macklin et al. 2006b; Turner et al. 2002, 2008; Turner 2004). On 25 April 1998, a tailings dam failed at the Frailes lead and zinc mine at Aznalcóllar near Seville in south-west Spain. The spill released over 4 million m3 of toxic slurries into the Río Agrio, a tributary of the Río Guadiamar. This accident covered several thousand hectares of farmland with contaminated water and sediment and has had a long-lasting impact on the Río Guadiamar that received most of the tailings waste and

346

Mark Macklin and Jamie Woodward

Fig. 11.19. The deeply incised channel zone of the Alfíos River in western Greece. Intensive in-stream gravel extraction has produced rapid channel degradation and bank retreat (photo: Jamie Woodward).

the internationally important wetlands of the Doñana National Park on the Atlantic coast (see Chapter 9). This UN World Heritage Area is still receiving significantly elevated concentrations of heavy metals ten years after the spill. Sediment-borne metals derived from the erosion and remobilization of contaminated river channel and floodplain sediment is the major problem at this site. During the course of the twentieth century, the sediment load reaching the Rhône delta declined by a factor of 4.7 from 35 to 7.39 million tonnes per year largely due to the construction of large dams in the catchment. Channel incision on the Rhône delta began in 1895, eventually leading to a lowering of the floodplain water table and salinization of floodplain soils as saline waters migrated upstream from the coastal zone (Arnaud-Fassetta 2003). In western Greece, the lower reaches of the Alfios River basin (3,600 km2 ) were transformed in less than three decades following intensive gravel extraction from the bed of the active channel (Figure 11.19). Between 1967 and 1995 more than 17 million m3 of gravel were extracted using large cranes and drag buckets (Christopoulos 1998; Nicholas et al. 1999). This activity took place in the reaches immediately downstream of

the Flokas Dam which is only about 5 km downstream of the archaeological site of Olympia—where this chapter began! The dam prevents bed load replenishment in the reaches downstream and this amplified the response of the channel to a major perturbation to the coarse sediment budget. This led to a period of rapid and sustained channel degradation resulting in 6–8 m of vertical incision and mean annual bank erosion rates of about 10 m per year with major bank failures associated with large flood events. Air photographs show that, in the gravel extraction reaches (immediately downstream of the Flokas Dam), the main channel of the Alfios River may have widened by up to 100 m between 1960 and 1990 (Christopoulos 1998; Nicholas et al. 1999). Alluvial gravel extraction is common throughout the Mediterranean region where gravel bed rivers with broad channel zones and seasonally low stages provide ready access to this important resource. The regulation of these gravel extraction activities is commonly ineffective and excessive rates of extraction can promote rapid channel changes that may propagate both upstream and downstream. This has led to falling water tables, elevated turbidity, and large tracts of valley floor with severely degraded channel and riparian habitats.

Rivers and Environmental Change

347

Fig. 11.20. Summary of human impacts on the river channel systems in the lower reaches and delta complex of the Axíos, Aliakmon, and Gallikos rivers in north-east Greece over the last century (modified from Kapsimalis et al. 2005). The darker the bars the greater the intensity of the human impact.

Fish species that spawn in gravel substrates are also threatened by gravel extraction. Cote et al. (1999) have shown how populations of river blennies (Salaria fluviatilis), that have a wide circum-Mediterranean distribution, are threatened by gravel extraction because structural alterations to gravel beds render the habitats unsuitable for breeding. In north-east Greece, Kapsimalis et al. (2005) have investigated recent changes in the lower reaches of three large rivers systems that drain to the Inner Thermaikos Gulf in the north-west Aegean Sea. The catchments of the Axios (23,747 km2 ), Aliakmon (9,250 km2 ), and Gallikos (1,230 km2 ) rivers accounted for a mean annual water discharge of approximately 276 m3 s−1 between 1926 and 1970, but this has declined in recent years to c.130 m3 s−1 because of abstractions for irrigation and urban consumption. Figure 11.20 provides a summary of the major changes between 1850 and 2000 to the lower reaches of these rivers following the increasing intensity of human impacts on water and sediment fluxes. The list of human activities shown in Figure 11.20 is typical of the impacts seen in many river basins across the Mediterranean region in the twentieth century (see Surian and Rinaldi 2003). These impacts have led to dramatic transformations to alluvial channels, floodplains, and deltaic environments. The net result of these activities has been a radical transformation of catchment water and sediment budgets and of

channel and floodplain environments across the region and these impacts are especially apparent in the more populous European Mediterranean.

Conclusions The Mediterranean basin provides a unique region in which to examine the long-term impact of environmental change on river system dynamics. The tectonic disposition of the region means that Quaternary fluvial sediments and landforms are commonly well preserved in river basins across the region and they represent a very significant archive of environmental change. The rich archaeological records have allowed human– river environment interactions to be studied in unusual detail from Palaeolithic to recent times. The interaction between fluvial geomorphologists and archaeologists over many decades has generated important ideas and debates about river system development that have resonated well beyond the Mediterranean world. Recent advances include the development of much more robust dating frameworks for fluvial sequences and a greater appreciation of the significance of large floods in the creation of the fluvial archive. However, there is still much work to do if we are to resolve outstanding issues associated with the investigation and interpretation of the Holocene record. More detailed records

348

Mark Macklin and Jamie Woodward

of land use change are needed in many areas. Closer collaboration at the basin scale between geoscientists and archaeologists and historians would improve our understanding of the impacts of land use change on the behaviour of river basin systems. The flood hydrology and sediment loads of many catchments across the Mediterranean were modified significantly by the climate changes associated with the Little Ice Age and data from this period provide a valuable guide for our interpretation of earlier parts of the fluvial archive. Researchers now have a much better appreciation of the role of large flood events in Mediterranean river systems through the systematic investigation of slackwater sediments, boulder flood units, documentary evidence, and the monitored hydrological record. Environmental changes associated with direct human impacts on river systems have intensified over the last 150 years or so and many valley floor environments in the Mediterranean have been degraded by resource exploitation, urban expansion, and the impacts of water and sediment pollution.

Acknowledgements This chapter was completed in the Acropole Hotel in Khartoum about 3,000 km upstream of the Mediterranean Sea. We would like to thank the external reviewer for a thoughtful review of the text and Nick Scarle for producing the diagrams. Our work in the Mediterranean has been supported by various bodies over the last two decades including the NERC and by many interdisciplinary archaeological projects. It has also involved collaboration with several generations of Ph.D. students and with many colleagues and good friends. We are grateful to them all.

References Abbott, J. T. and Valestro, S. (1995), The Holocene alluvial records of the chorai of Metapontum, Basilicata, and Croton, Calabria, Italy, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 195–205. Allen, J. R. M., Brandt, U., Brauer, A., Hubberten, H. W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J. F. W., Nowaczyk, N. R., Oberhansli, H., Watts, W. A., Wulf, S., and Zolitschka, B. (1999), Rapid environmental changes in southern Europe during the last glacial period. Nature 400: 740–3. Almagià, R. (1918), Trace glaciali nelle montagne dell’Albania. Rivista Geografica Italiana 25: 85–95. Anketell, J. M., Ghellali, S. M., Gilbertson, D. D., and Hunt, C. O. (1995), Quaternary wadi and floodplain sequences of Tripolitania, northwest Libya: A synthesis of their stratigraphic relationships, and their implications in landscape evolution, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 231–44.

Arnaud-Fassetta, G. and Provansal M., (1999), High frequency variations of water flux and sediment discharge during the Little Ice Age (1586–1725 AD) in the Rhône Delta (Mediterranean France). Relationship to the catchment basin. Hydrobiologia 410: 241–50. (2003), River channel changes in the Rhône Delta (France) since the end of the Little Ice Age: geomorphological adjustment to hydroclimatic change and natural resource management. Catena 51: 141–72. Bailey, G. N. (ed.) (1997), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, i. Excavation and Intrasite Analysis at Klithi. McDonald Institute for Archaeological Research, Cambridge. Lewin, J., Macklin, M. G., and Woodward, J. C. (1990), The ‘Older Fill’ of the Voidomatis Valley Northwest Greece and its relationship to the Palaeolithic Archaeology and Glacial History of the Region. Journal of Archaeological Science 17: 145–50. Barriendos Vallve, M. and Martin-Vide, J. (1998), Secular Climatic Oscillations as Indicated by Catastrophic Floods in the Spanish Mediterranean Coastal Area (14th–19th Centuries). Climatic Change 38: 473–91. Bavec, M., Tulaczyk, S. M., Mahan, S. A., and Stock, G. M. (2004), Late Quaternary glaciation of the upper Soˇca River region, southern Julian Alps, NW Slovenia. Sedimentary Geology 165: 265–83. Benito, G. (2003), Palaeohydrological changes in the Mediterranean region during the Late Quaternary, in K. J. Gregory and G. Benito (eds.), Palaeohydrology: Understanding Global Change. John Wiley, Chichester, 123–42. Sopeñab, A., Sánchez-Moyac, Y., Machadod, M. J., and Pérez-González, A. (2003a ), Palaeoflood record of the Tagus River (Central Spain) during the Late Pleistocene and Holocene. Quaternary Science Reviews 22: 1737–56. Díez-Herrero, A., and Fernández de Villalta, M. (2003b), Magnitude and frequency of flooding in the Tagus basin (central Spain) over the last millennium. Climatic Change 58: 171–92. Bintliff, J. L. (1992), Erosion in the Mediterranean lands: A reconsideration of pattern, process and methodology, in M. G. Bell and J. Boardman (eds.), Past and Present Soil Erosion. Oxbow, Oxford, 125–31. (2002), Time, process and catastrophism in the study of Mediterranean alluvial history: A review. World Archaeology 33: 417–35. and Snodgrass, A. M. (1985), The Cambridge/Bradford Boeotian Expedition: The first four years. Journal of Field Archaeology 12: 123–61. Blissenbach, E. (1954), Geology of alluvial fans in semiarid regions. Geological Society of America Bulletin 65: 175–89. Bond, G., Broecker, W., Johnsen, S., McManus, J., Labeyrie, L., Jouzel, J., and Bonani, G. (1993), Correlations between climate records from North Atlantic sediments and Greenland ice. Nature 365: 143–7. Kromer, B., Beer, J., Muscheler, R., Evans, M. N., Showers, W., Hoffman, S., Lotti-Bond, R., Hajdas, I., and Bonani, G. (2001), Persistent solar influence on North Atlantic climate during the Holocene. Science 294: 2130–6. Branca, S. and Ferrara, V. (2001), An example of river pattern evolution produced during the lateral growth of a central polygenic volcano: the case of the Alcantara river system, Mt Etna (Italy). Catena 45: 85–102.

Rivers and Environmental Change Brückner, H. (1986), Man’s impact on the evolution of the physical environment in the Mediterranean region in historical times. GeoJournal 13: 7–17. Butzer, K. W. (1969), Changes in the land. Science 165: 52–3. Butzer, K. W. (1977), Geo-archaeology in practice. Reviews of Anthropology 4: 125–31. (1980), Holocene alluvial sequences: problems of dating and correlation, in R. A. Cullingford, D. A. Davidson, and J. Lewin (eds.), Timescales in Geomorphology. John Wiley & Sons, Chichester, 131–42. Candy, I., Black, S., and Sellwood, B. W. (2005), U-series isochron dating of immature and mature calcretes as a basis for constructing Quaternary landform chronologies for the Sorbas basin, southeast Spain. Quaternary Research 64: 100–1. Carrión, J. S. (2002), Patterns and processes of Late Quaternary environmental change in a montane region of southwestern Europe. Quaternary Science Reviews 21: 2047–66. Casford, J. S. L., Abu-Zied, R. H., Rohling, E. J., Cooke, S., Boessenkool, K. P., Brinkhuis, H., de Vries, C., Wefer, G., Geraga, M., Papatheodorou, G., Croudace, I., Thomson, J., Wells, N. C., and Lykousis, V. (2001), Mediterranean climate variability during the Holocene. Mediterranean Marine Science 2: 45–55. Christopoulos, G. (1998), Late Holocene river behaviour in the lower Alfios basin, western Greece. Ph.D. Thesis, University of Leeds. Cviji´c, J. (1900), L’Époque Glaciaire dans la Péninsule des Balkans. Annales de Géographi 9: 359–72. Collier, R. E. L., Leeder, M. R., and Jackson, J. A. (1995), Quaternary drainage development, sediment fluxes and extensional tectonics in Greece, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 31–44. Conchon, O. (1986), Quaternary glaciations in Corsica. Quaternary Science Reviews 5: 429–32. Cote, I. M., Vinyoles, D., Reynolds, J. D., Doadrio, I., and Perdices, A. (1999), Potential impacts of gravel extraction on Spanish populations of river blennies Salaria fluviatilis (Pisces, Blenniidae). Biological Conservation 87: 359–67. Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjörnsdottir, A. E., Jouzel, J., and Bond, G. (1993), Evidence for general instability of past climate from a 250-kyr ice-core record. Nature 364: 218–20. Davidson, D. A. and Shackley, M. R. (1976), Geoarchaeology: Earth Science and the Past. Duckworth, London. Dewey, J. F., Pitman III, W. C., Ryan, W. B. F., and Bonnin, J. (1973), Plate tectonics and the evolution of the Alpine system. Geological Society of America Bulletin 84: 3137–80. Fuller, I. C., Macklin, M. G., Passmore, D. G., Brewer, P. A., Lewin, J., and Wintle, A. G. (1996), Geochronologies and environmental records of Quaternary fluvial sequences in the Guadalope basin, northeast Spain, based on luminescence dating, in J. Branson, A. G. Brown, and K. J. Gregory (eds.), Global Continental Changes: the Context of Palaeohydrology. Geological Society of London Special Publication 115: 99–120. (1998), River response to highfrequency climate oscillations in southern Europe over the past 200 k.y. Geology 26: 275–8. Gamble, C. S. (1986), The Palaeolithic Settlement of Europe. Cambridge University Press, Cambridge. Gaucher, G. (1947), Les dépôts quaternaires du Bas-Chélif et des basses plaines oranaises. CR Acad. Sci. Paris 225: 65–6.

349

Gigout, M. (1959), A propos du Quaternaire marin sur le litoral de la Provincia de Tarragona. Boletin. Com. Mapa. Geol. Tomo, IV Madrid. Gilbertson, D. D., Hunt, C. O., and Smithson, P. A. (1996), Quaternary geomorphology and palaeoecology, in G. W. W. Barker, D. D. Gilbertson, G. D. B. Jones, and D. J. Mattingly (eds.), Farming the Desert: The UNESCO Libyan Valleys Archaeological Survey i: 49–82. Greenbaum, N., Asher, P., Schick, A. P., and Baker, V. R. (2000) The palaeoflood record of a hyperarid catchment, Nahal Zin, Negev Desert, Israel. Earth Surface Processes and Landforms 25: 951–71. Porat, N., Rhodes, E., and Enzele, Y. (2006) Large floods during late Oxygen Isotope Stage 3, southern Negev desert, Israel. Quaternary Science Reviews 25: 704–19. Griffiths, J. S., Mather, A. E., and Hart, A. B. (2002), Landslide susceptibility in the Río Aguas catchment, SE Spain. Quarterly Journal of Engineering Geology and Hydrogeology 35: 9–17. Grove, A. T. (2001), The ‘Little Ice Age’ and its geomorphological consequences in Mediterranean Europe. Climatic Change 48: 121–36. Grove, J. M. (1988), The Little Ice Age. Routledge, London. and Conterio, A. (1995), The climate of Crete in the sixteenth and seventeenth centuries. Climatic Change 30: 223–47. Hamlin, R. H. B. (2000), Environmental change and catastrophic flooding in the Voidomatis and Aoos Basins, Northwes t Greece. Ph.D. Thesis, University of Leeds. Woodward, J. C., Black, S., and Macklin, M. G. (2000), Sediment fingerprinting as a tool for interpreting long-term river activity: the Voidomatis basin, NW Greece, in I. D. L. Foster (ed.), Tracers in Geomorphology. John Wiley & Sons, Chichester, 473–501. Harvey, A. M. and Wells, S. G. (1987), Response of Quaternary fluvial systems to differential epeirogenic uplift: Aguas and Feos river systems, southeast Spain. Geology 15: 689–93. Miller, S. Y., and Wells, S. G. (1995), Quaternary soil and river terrace sequences in the Aguas/Feos river systems: Sorbas basin, southeast Spain, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 263–81. Hewitt, S. (2002), Pleistocene river behaviour and sediment transfer in glaciated catchments of the Mediterranean. Ph.D. Thesis, University of Leeds. Hey, R. W. (1962), The Quaternary and Palaeolithic of Northern Libya, Quaternaria 6: 435–49. Higgs, E. S. and Vita-Finzi, C. (1966), The Climate, Environment and Industries of Stone Age Greece: Part II. Proceedings of the Prehistoric Society 32: 1–29. Hudson-Edwards, K. A., Macklin, M. G., Jamieson, H. E., Brewer, P. A., Coulthard, T. J., Howard, A. J., and Turner, J. N. (2003), The impact of tailings dam spills and clean-up operations on sediment and water quality in river systems: The Ríos Agrio-Guadiamar, Aznalcóllar, Spain. Applied Geochemistry 18: 221–39. Hughes, P. D., Woodward, J. C., and Gibbard, P. L. (2006a ), Glacial history of the Mediterranean mountains. Progress in Physical Geography 30: 334–64. Macklin, M. G., Gilmour, M. A., and Smith, G. R. (2006b), The glacial history of the Pindus Mountains, Greece. Journal of Geology 114: 413–34. Huntington, E. (1910), The Burial of Olympia. Geographical Journal 36: 657–86.

350

Mark Macklin and Jamie Woodward

Judson, S. (1963), Erosion and deposition of Italian stream valleys during historical time. Science 140: 898–9. Kapsimalis, V., Poulos, S. E., Karageorgis, A. P., Pavlakis, P., and Collins, M. (2005), Recent evolution of a Mediterranean deltaic coastal zone: human impacts on the Inner Thermaikos Gulf, NW Aegean Sea. Journal of the Geological Society 162: 897–908. Kelly, M., Black, S., and Rowan, J. S. (2000), A calcrete-based U/Th chronology for landform evolution in the Sorbas basin, Southeast Spain. Quaternary Science Reviews 19: 995–1010. Kraft, J. C., Rapp, G., and Aschenbrenner, S. E. (1975), Holocene palaeogeography of the coastal plain of the Gulf of Messenia, Greece, and its relationship to archaeological setting and coastal change. Bulletin of the Geological Society of America 86: 1191–208. Kayan, I., and Erol, O. (1980), Geomorphic Reconstructions in the Environs of Ancient Troy. Science. 209: 776–82. Bruckner, H., Kayan, I., and Engelman, H. (2007), The geographies of Ancient Ephesus and the Artemision in Anatolia. Geoarchaeology: An International Journal 22: 121–49. Lahlou, A. (1988), The silting of Moroccan dams. In: M. P. Bordas and D. E. Walling (eds.), Sediment Budgets. IAHS Publication No. 174: 71–77. Lewin, J. and Macklin, M. G. (2003), Preservation potential for Late Quaternary river alluvium. Journal of Quaternary Science 18: 107–20. and Woodward, J. C. (1991), Late Quaternary fluvial sedimentation in the Voidomatis Basin, Epirus, northwest Greece. Quaternary Research 35: 103–15. (eds.) (1995), Mediterranean Quaternary River Environments. A. A. Balkema, Rotterdam. López-Avilés, A. (1999), Flood regime and Quaternary sedimentation style in the Guadalope Basin, northeast Spain. Ph.D. Thesis University of Leeds. Ashworth, P. J., and Macklin, M. G. (1998), Floods and Quaternary sedimentation style in a bedrock-controlled reach of the Bergantes River, Ebro Basin, northeast Spain, in G. Benito, V. R. Baker, and K. J. Gregory (eds.), Palaeohydrology and Environmental Change. Wiley, Chichester, 181–96. Lowe, J. J. and Walker, M. J. C. (1997), Reconstructing Quaternary Environments (2nd edn.). Addison-Wesley-Longman, London. Maas, G. S. (1998), River response to Late Quaternary environmental change, southwestern Crete, Greece. Ph.D. thesis University of Leeds. and Macklin, M. G. (2002), The impact of recent climate change on flooding and sediment supply within a Mediterranean mountain catchment, southwestern Crete, Greece. Earth Surface Processes and Landforms 27: 1087–105. McBurney, C. B. M. and Hey, R. W. (1955), Prehistory and Pleistocene Geology in Cyrenaican Libya. A Record of Two Seasons’ Geological and Archaeological Fieldwork in the Gebel Akhdar Hills, with a Summary of Prehistoric Finds from Neighbouring Territories. Cambridge University Press, Cambridge. Macklin, M. G. (1995), Geochronology, correlation and controls of Quaternary river erosion and sedimentation, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 179–81. and Lewin, J. (2003), River sediments, great floods and centennial-scale Holocene climate change. Journal of Quaternary Science 18: 102–5.

and Needham, S. (1992), Studies in alluvial archaeology: potential and prospect, in S. Needham and M. G. Macklin (eds.), Alluvial Archaeology in Britain. Oxbow, Oxford, 9–23. and Passmore, D. G. (1995), Pleistocene environmental change in the Guadalope basin, northeast Spain: fluvial and archaeological records, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Rotterdam, Balkema, 103–113. Lewin, J., and Woodward, J. C. (1995), Quaternary fluvial systems in the Mediterranean basin, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 1–25. (1997), Quaternary River Sedimentary Sequences of the Voidomatis Basin, in G. N. Bailey (ed.), Klithi: Palaeolithic Settlement and Quaternary Landscapes in Northwest Greece, ii. Klithi in its Local and Regional Setting. McDonald Institute for Archaeological Research, Cambridge, 347–59. Fuller, I. C., Lewin, J., Maas, G. S., Passmore, D. G., Rose, J., Woodward, J. C., Black, S., Hamlin, R. H. B., and Rowan, J. S. (2002), Correlation of Late and Middle Pleistocene fluvial sequences in the Mediterranean and their relationship to climate change. Quaternary Science Reviews 21: 1633–44. Johnstone, E., and Lewin, J. (2005), Pervasive and longterm forcing of Holocene river instability and flooding by centennial-scale climate change. The Holocene 15: 937–43. Benito, G., Gregory, K. J., Johnstone, E., Lewin, J., ´ Michczynska, D. J., Soja, R., Starkel, L., and Thorndycraft, V. R. (2006a ), Past hydrological events reflected in the Holocene fluvial record of Europe. Catena 66: 145–54. Brewer, P. A., Hudson-Edwards, K. A., Bird, G., Coulthard, T. J., Dennis, I. A., Lechler, P. J., Miller, J. R., and Turner, J. N. (2006b), A geomorphological approach to the management of rivers contaminated by metal mining. Geomorphology 79: 423–47. McManus, J. F., Anderson, R. F., Broecker, W. S., Fleisher, M. Q., and Higgins, S. M. (1998), Radiometrically determined sedimentary fluxes in the sub-polar North Atlantic during the last 140,000 years. Earth and Planetary Science Letters 155: 29–43. Magny, M., Bégeot, C., Guiot, J., and Peyron, O. (2003), Contrasting patterns of hydrological changes in Europe in response to Holocene climate cooling phases. Quaternary Science Reviews 22: 1589–96. Mather, A. E. (2000), Adjustment of a drainage network to capture induced base-level change. Geomorphology 34: 271–89. Harvey, A. M., and Brencheley, P. J. (1991), Halokinetic deformation of Quaternary river terraces in the Sorbas basin, SE Spain. Zeitfrich für Geomorphologie Suppl. 82: 87–97. Moody, J. and Grove, A. T. (1990), Terraces and enclosure walls in the Cretan landscape, in S. Bottema, G. Entjes-Nieborg, and W. van Zeist (eds.), Man’s Role in the Shaping of the Eastern Mediterranean Landscape. A. A. Balkema, Rotterdam, 183–91. Nicholas, A. P., Woodward, J. C., Christopoulos, G., and Macklin, M. G. (1999), Modelling and monitoring the impact of dam construction and gravel extraction on rates of bank erosion in the Alfios River, Peloponnese, western Greece, in A. G. Brown and T. A. Quine (eds.), Fluvial Processes and Environmental Change. John Wiley & Sons, Chichester, 117–37. Niculescu, C. (1915), Sur les traces de glaciation dans le massif Smolica chaîne du Pinde méridional. Bulletin de la Section Scientifique de l’Academie Roumaine 3: 146–51.

Rivers and Environmental Change Noble, J. (2004), Response of a steepland river system to late Quaternary environmental change: south central Crete. Ph.D. Thesis, University of Wales, Aberystwyth. Paepe, R., Hatziotis, M. E., and Thorez, J. (1980), Geomorphological Evolution in the Eastern. Mediterranean Belt and Mesopotamian Plain. Report for the International Geological Correlation Programme Project 146: River flood and lake level changes. Palanques, A., Plana, F., and Maldonado, A. (1990), Recent influence of man on the Ebro margin sedimentation system, northwestern Mediterranean Sea. Marine Geology 95: 247–63. Pichard, G. (1995), Les crues sur le bas Rhône de 1500 à nos jours. Pour une histoire hydro-climatique. Méditerranée 3/4: 105–16. Piegay, H. and Bravard, J.-P. (1997), Response of a Mediterranean riparian forest to a 1 in 400 year flood, Ouveze River, Drome-Vaucluse, France. Earth Surface Processes and Landforms 22: 31–43. Pope, K. O. and van Andel, T. H. (1984), Late Quaternary alluviation and soil formation in the southern Argolid: its history, causes and archaeological implications. Journal of Archaeological Science 11: 281–306. Raphael, C. N. (1973), Late Quaternary Changes in Coastal Elis, Greece. Geographical Review 63: 73–89. Renfrew, A. C. and Wagstaff, M. (1982), An Island Polity: The Archaeology of Exploitation in Melos. Cambridge University Press, Cambridge. Roberts, N., Reed, J., Leng, M. J., Kuzucuo˘glu, C., Fontugne, M., Bertaux, J., Woldring, H., Bottema, S., Black, S., Hunt, E., and Karabıyıko˘glu, M. (2001), The tempo of Holocene climatic change in the eastern Mediterranean region: new highresolution crater-lake sediment data from central Turkey. The Holocene 11: 721–36. Rose, J. and Meng, X. (1999), River activity in small catchments over the last 140 ka, northeast Mallorca, Spain, in: A. G. Brown and T. A. Quine (eds.), Fluvial Processes and Environmental Change. Wiley, Chichester, 91–102. Roucoux, K. H., de Abreu, L., Shackleton, N. J., Tzedakis, P. C., (2005), The response of NW Iberian vegetation to North Atlantic climate oscillations during the last 65 kyr. Quaternary Science Reviews 25: 1637–53. Rowan, J. S., Black, S., Macklin, M. G., Tabner, B. J., and Dore, J. (2000), Quaternary environmental change in Cyrenaica evidenced by U-Th, ESR and OSL of coastal alluvial fan sequences. Libyan Studies 31: 5–16. Rumsby, B. A. and Macklin, M. G. (1996), River response to the last neoglacial (the ‘Little Ice Age’) in northern, western and central Europe, in K. J. Gregory (ed.), Global Continental Changes: The Context of Palaeohydrology. Geological Society Special Publication, London 115: 217–33. Sadori, L. (2001), The Postglacial record of environmental history from Lago di Pergusa, Sicily. The Holocene 11: 655–71. Sancho, C., Peña Monné, J. L., Lewis, C., McDonald, E., and Rhodes, E. (2003), Preliminary dating of glacial and fluvial deposits in the Cinca River Valley (NE Spain): chronological evidences for the Glacial Maximum in the Pyrenees? in B. Ruíz-Zapata, M. Dorado-Valiño, A. Valdemillos, M. J. Gil-García, T. Badají, I. Bustamante, I. Mendizábal (eds.), Quaternary Climatic Changes and Environmental Crises in the Mediterranean Region. Universidad de Alcalá de Henares, Madrid, 453–6. Santisteban, J. I. and Schulte, L. (2007), Fluvial networks of the Iberian Peninsula: a chronological framework. Quaternary Science Reviews 26: 2738–57.

351

Schulte, L., Julià, R., Burjachs, F., and Hilgers, A. (2007), Middle Pleistocene to Holocene geochronology of the River Aguas terrace sequence (Iberian Peninsula): fluvial response to Mediterranean environmental change. Geomorphology 98: 13–33. Selli, R. (1962), Le Quaternaire marin du versant Adriatique Ionien de la péninsule italienne. Quaternaria 6: 391–413. Surian, N. (1999), Channel changes due to river regulation: the case of the Piave River, Italy. Earth Surface Processes and Landforms 24: 1135–51. and Rinaldi, M. (2003), Morphological response to river engineering and management in alluvial channels in Italy. Geomorphology 50: 307–26. Thorndycraft, V. R. and Benito, G. (2006), Late Holocene fluvial chronology of Spain: The role of climatic variability and human impact. Catena 66: 34–41. Turner, J. N. (2004), Ageomorphological-geochemical assessment of the impacts of the April 1998 Aznalcóllar tailings dam failure on the Rio Guadiamar, southwest Spain. Ph.D. Thesis, University of Wales, Aberystwyth. Brewer, P., Macklin, M. G., Hudson-Edwards, K. A., Coulthard, T. J., Howard, A. J., and Jamieson, H. E. (2002), Heavy metal and as transport under low and high flows in the River Guadiamar three years after the Aznalcollar tailings dam failure: Implications for river recovery and management, in J. M. Garcia, J. A. A. Jones, and J. Arnaez (eds.), Environmental Change and Water Sustainability. IPE-CSIC, Zaragoza, 235–51. (2008) Fluvial-controlled metal and as mobilisation, dispersal and storage in the Río Guadiamar, SW Spain and its implications for long-term contaminant fluxes to the Doñana wetlands. Science of The Total Environment 394: 144– 61. Tzedakis, P. C., Lawson, I. T., Frogley, M. R., Hewitt, G. M., and Preece, R. C. (2002), Buffered tree population changes in a Quaternary refugium: evolutionary implications. Science 297: 2044–7. van Andel, T. H. and Runnels, C. (1987), Beyond the Acropolis: A Rural Greek Past. Cambridge University Press, Cambridge. Runnels, C. N., and Pope, K. O. (1986), Five thousand years of land use and abuse in the southern Argolid. Hesperia 55: 103–28. Zangger, E., and Demitrack, A. (1990), Land use and soil erosion in Prehistoric and Historical Greece. Journal of Field Archaeology 17: 379–96. Vita-Finzi, C. (1969), The Mediterranean Valleys: Geological Changes in Historical Times. Cambridge University Press, Cambridge. (1976), Diachronism in Old World alluvial sequences. Nature 263: 218–19. (1986), Recent Earth Movements: An Introduction to Neotectonics. Academic Press, London. Westaway, R., Pringle, M., Yurtmen, S., Demir, T., Bridgland, D., Rowbotham, G., and Maddy, D. (2003), Pliocene and Quaternary surface uplift of western Turkey revealed by long-term river terrace sequences. Current Science 84: 1090–101. (2004), Pliocene and Quaternary regional uplift in western Turkey: the Gediz River terrace staircase and the volcanism at Kula. Tectonophysics 391: 121–69.

352

Mark Macklin and Jamie Woodward

Woodward, J. C. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water Quality in River Catchments. John Wiley & Sons Chichester, 365–89. and Foster, I. D. L. (1997), Erosion and suspended sediment transfer in river catchments: environmental controls, processes and problems. Geography 82/4: 353–76. Lewin, J., and Macklin, M. G. (1992), Alluvial sediment sources in a glaciated catchment: the Voidomatis basin, northwest Greece. Earth Surface Processes and Landforms 16: 205–16. Macklin, M. G., and Lewin, J. (1994), Pedogenic weathering and relative age dating of Quaternary alluvial sediments in the Pindus Mountains of northwest Greece, in D. A. Robinson and R. B. G. Williams (eds.), Rock Weathering and Landform Evolution. John Wiley & Sons, Chichester, 259–83. Lewin, J., and Macklin, M. G. (1995), Glaciation, river behaviour and the Palaeolithic settlement of upland northwest Greece, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 115–29.

Hamlin, R. H. B., Macklin, M. G., Karkanas, P., and Kotjabopoulou, E. (2001), Quantitative sourcing of slackwater deposits at Boila Rockshelter: A record of Lateglacial flooding and Palaeolithic settlement in the Pindus Mountains, Northwest Greece. Geoarchaeology: An International Journal 16: 501–36. Macklin, M. G., and Smith, G. R. (2004), Pleistocene glaciation in the mountains of Greece, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations—Extent and Chronology: Part 1. Elsevier, Amsterdam, 155–73. Hamlin, R. H. B., Macklin, M. G., Hughes, P. D., and Lewin, J. (2008), Glacial activity and catchment dynamics in northwest Greece: Long-term river behaviour and the slackwater sediment record for the last glacial to interglacial transition. Geomorphology 101: 44–67. Zielhofer, C., Faust, D., Baena Escudero, R., Diaz del Olmo, F., Kadereit, A., Moldenhauer, K., and Porras, A. (2004), Centennial-scale Late Pleistocene to mid-Holocene synthetic profile of the Medjerda Valley, northern Tunisia. The Holocene 14: 851–61. Zorzou, M. (2004), Suspended sediment delivery and sediment properties in mountain catchments of western Greece. Ph.D. thesis, University of Leeds.

This chapter should be cited as follows Macklin, M. G. and Woodward, J. C. (2009) River systems and environmental change, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 319–352.

12

Glacial and Periglacial Environments Philip Hughes and Jamie Woodward

Introduction Traditionally, glacial and periglacial geomorphology has not featured prominently in discussions about the physical geography of the Mediterranean basin. It is now clear, however, that on numerous occasions during the Pleistocene, and to a lesser extent during the Little Ice Age (LIA), glacial and periglacial activity was widespread in many of the region’s mountain ranges (Hughes et al. 2006a; Hughes and Woodward 2008). Even today, small glaciers and active periglacial features can be found on the highest peaks. Many mountain landscapes in the Mediterranean basin are therefore the product of glacial and periglacial processes that have fluctuated in intensity and spatial extent through the Quaternary. Glacial processes are defined here as those occurring as a result of dynamic glacier ice. The periglacial zone is sometimes defined as non-glacial areas where the mean annual temperature is less than 3◦ C (French 1996: 20). However, cryogenic processes can be important in landform development, even in areas of shallow frost over a wide range of mean annual temperatures. Thus, the term ‘periglacial’ is applied here to areas characterized by cold-climate processes—where frost and nival processes are important—but where glaciers are absent. Glacial and periglacial processes in the uplands can exert considerable influence upon geomorphological systems at lower elevations. Fluvial systems, for example, over a range of timescales have been shown to be especially sensitive to changes in sediment supply and water discharge from glaciated mountain headwaters (Gurnell and Clark 1987; Woodward et al. 2008). Nonetheless, the geomorphological impacts of glaciation

are most clearly evident in the Mediterranean mountains where the erosional and depositional legacy is frequently well preserved. Cirques, glacial lakes, icescoured valleys, moraines, pronival ramparts, relict rock glaciers, and other glacial and periglacial features can be found in many Mediterranean mountain ranges (Hughes et al. 2006a). Upland limestone terrains are widespread across the Mediterranean and many of these landscapes have been shaped by a combination of glacial and karstic processes (Chapter 10). In fact, glacio-karst is probably the dominant landscape in many mountain regions, including the Dinaric Alps of Croatia/Bosnia/Montenegro (Nicod 1968), the Cantabrian Mountains of Spain (Smart 1986) and the Pindus Mountains of Greece (Waltham 1978; Woodward et al. 2004; Hughes et al. 2006b). Glaciated mountain landscapes are also found in other rock types including granite (e.g. Monte Cinto, Corsica) (Conchon 1978) and ophiolite (e.g. Mount Smolikas, Greece) (Hughes et al.2006c). Glacial and periglacial landforms provide an important record of environmental change and can be used to reconstruct past glacial climates. This chapter examines the evidence for both past and present glacial and periglacial activity in the Mediterranean mountains and assesses its wider significance in the region. All the sites referred to in this chapter are shown in Figure 12.1 and all elevations are given in metres above sea level. This chapter is divided into three main parts. These consider, in turn, glacial and periglacial environments of the present, during the Holocene (with emphasis on the LIA), and during the cold stages of the Pleistocene.

354

Philip Hughes and Jamie Woodward

13 2 1

3

4 6

5

9

8

14 10

11

18

7

25 26

12

19 17

11

Currently glacierized areas

1

Other mountain areas

0

Serra da Estrela Serra do Gêres Serra de Queixa Cantabrian Mountains Sierra de Gredos Peñalara Massif

7 8 9 10 11 12

21

20

24

27

1 2 3 4 5 6

22

16 15

Sierra Nevada Pyrenees Alpes Maritimes Corsica Italian Apennines Mount Etna

13 14 15 16 17 18

23

1000 km

Julian Alps Dinaric Alps Pindus Mountains Mount Olympus Crete Pirin/Rila Mountains

19 20 21 22 23 24

Western Taurus Central Taurus Eastern Taurus Pontic Mountains Jbel Liban Aurès Massif

25 Djurdjura Massif 26 Middle Atlas 27 High Atlas

Fig. 12.1. Map of the Mediterranean showing the main mountain areas referred to in this chapter. Currently glacierized areas are indicated.

Modern Glacial and Periglacial Environments Modern glaciers are rare in the Mediterranean because most mountains do not reach the snowline. Glaciers are present, however, under modern climatic conditions in the Pyrenees, Alpes Maritimes, Italian Apennines, Julian Alps, and in the mountains of Montenegro and Turkey (Table 12.1 and Figure 12.1). These glaciers are mainly small cirque glaciers on mountains above 2,500 m, but extensive valley glaciers and small ice caps are present on some of the highest mountains (>4,000 m). Periglacial processes are much more widespread and are active above 2,000 m in areas as far south as the Moroccan Atlas and at lower elevations in the northern Mediterranean. Snowfall is heavy in many of the Mediterranean mountains with most falling during the winter months between November and March (Chapter 3).

Turkey Thirty-eight modern glaciers have been identified in the mountains of Turkey and the largest Mediterranean

TABLE 12.1. Modern glaciers in the Mediterranean Mountain area

Pyrenees Alpes Maritimes Apennines Julian Alps Turkey Montenegro

Number of glaciers 41 6 1 1 40+ 1

Total glacier area (km2 )

ELA (m)

11.43 0.31 0.045 0.0303 22.9 0.05

2,320–3,005 2,800 2,750 2,500 2,900–4,100 2,150

Maximum glacier length (km) 2.0 <0.5 0.4 <0.5 4.0 <0.5

Sources: Based on data provided in Serrat and Ventura (1993), J. M. Grove (2004), Gellatly et al. (1994), Gabrovec (1998), Çiner (2004), Kurter and Sungur (1980), and Hughes (2007).

glaciers occur in the mountains of eastern Turkey (Akçar and Schlüchter 2005). The presence of glaciers in these mountains was noted in the nineteenth century by Ainsworth (1842) and Palgrave (1872). In 1980, glaciers covered a total area of c.22.9 km2 (Kurter and Sungur 1980), although almost everywhere glaciers are in retreat (Çiner 2004). Periglacial activity is also

Glacial and Periglacial Environments

likely to be a major influence on modern landform development in these areas, but its effects have not been studied in detail. Mount Ararat (A˘gri Da˘gi), a dormant stratovolcano situated in the easternmost Taurus Mountains, is the highest peak in Turkey (5,137 m) and is covered by an ice cap of about 10 km2 . This represents the largest single ice mass on any Mediterranean mountain. The ice cap extends down to c.4,100 m and has a snowline altitude of c.4,300 m (Kurter and Sungur 1980). Glaciers are found on other Turkish volcanoes, including Mount Süphan (4,058 m) and Mount Erciyes (3,917 m). On Mount Süphan, several glaciers occur in the crater, the largest being on a north-facing slope with a length of 1.5 km (Kurter 1991; Çiner 2004). The glacier on Mount Erciyes has been the focus of studies throughout the twentieth century and is known to have retreated from a length of 700 m in 1905 to a length of only 380 m by 1983 (Sarikaya 2003; Çiner 2004). These three volcanoes are now dormant, but K-Ar isotope age dating of lavas indicate volcanic activity during the Quaternary (Yilmaz et al. 1998). Volcanic activity is therefore likely to have interacted with glacial ice, especially during cold stages when glaciers were much more extensive. Three glacial stratigraphical units have been identified on Mount Ararat and two of these are covered by volcanic deposits (ibid.), although the relationship between dated volcanic units and glacial units on Mount Ararat is not clear. The south-eastern Taurus range (Güneydo˘gu Toroslar) contains the greatest concentration (n = 20) of modern glaciers in Turkey and the wider Mediterranean region (Figure 12.1). These are also the most southerly glaciers in the Mediterranean basin and owe their existence to the high altitude of these mountains (>4,000 m) (Kurter 1991). The largest occur on Mount Cilo (4,135 m) where the Uludoruk glacier is nearly 4 km long and covers an area of 8 km2 . In common with many other glaciers in Turkey, it has retreated throughout the twentieth century and the altitude of the glacier front rose approximately 400 m between 1937 and 1991 (Çiner 2004). Smaller glaciers occur in the central Taurus (Orta Toroslar), further to the west on the mountains of Alada˘g (3,756 m) and Bolkarda˘g (3,524 m). The snowline in these areas is situated at c.3,450 m and glacier survival in such marginal conditions is controlled by local climatological and physiographic conditions (Kurter 1991). In the Alada˘g Massif, a small glacier covering an area of <1 km2 is considered to be a glacier remnant consisting largely of dead ice buried by rock debris. A considerable thickness (c.120 m) of perennial

355

snow, firn, and ice has been reported from this site in a deep karstic shaft at c.3,400 m (Bayari et al. 2003). The northern slopes of the Pontic Mountains drain into the Black Sea and receive some of the highest precipitation levels in Turkey. In the east of this area, for example, mean annual precipitation exceeds 2,000 mm (Kurter 1991). As a result, the snowline on the northern slopes is between 3,100 and 3,400 m and this is considerably lower than on most other mountains in Turkey. At least twelve small glaciers exist in the Pontic Mountains and, together, they cover an area of about 2.54 km2 (Çiner 2004).

The Balkans Modern glaciers are not present in Greece since the permanent snowline is situated well above most of the highest peaks. The snowline is above 3,000 m in the Pindus Mountains of northern Greece and above 3,500 m across Crete (Messerli 1980). Nevertheless, all these mountains experience heavy winter snowfalls and, on the highest peaks (>2,500 m) such as Mount Olympus (2,912 m), even summer snowfall is common (Sahsamanoglu 1989). On the highest slopes of the Pindus Mountains periglacial features are very well developed and some appear to be active today. Thick accumulations of snow undergo cycles of freezing and thawing during the spring and summer melt. Patterned ground, solifluction, and related features occur in the zone above 1,800 m as far south as Crete, and have been attributed to presentday periglacial processes (Poser 1957). Periglacial activity is likely to be a significant geomorphological agent in the highest areas of the Greek mountains where mean monthly temperatures are well below −5◦ C and mean minimum temperatures are between −20 and −25◦ C during the coldest month (Furlan 1977: 195, 201). Lapse rate data indicate that the mean annual temperatures at 2,400 m are likely to be <3◦ C, so the highest areas of the Pindus Mountains can be considered ‘true’ periglacial environments, as defined by French (1996: 23). Further north in the former Yugoslavian republics and in Albania, the climate is harsher and more continental than in Greece. An absolute minimum winter temperature of −29.6◦ C has been observed at the summit of Bjelašnica (2,067 m) in central Bosnia and mean temperatures in January are below −5◦ C over large areas of the mountainous Balkan interior (Furlan 1977). Snowfalls are heavy in most areas; with drifts locally exceeding several metres (Figure 12.2a) and snow can persist well into summer (Figure 12.2b).

356

Philip Hughes and Jamie Woodward

(a)

(b) Fig. 12.2. (a) Snow accumulation to a depth of c.3 m in early June at an altitude of c.2,000 m a.s.l. between the villages of Zabljak and Crna Gora on the northern slopes of Durmitor (2,522 m a.s.l.), in Montenegro. (b) Looking west from the summit of Mount Orjen (1,894 m a.s.l.) towards the subsidiary summit of Subra (1,679 m a.s.l.), which is situated less than 15 km from the coast of Montenegro. In this area, annual precipitation values average over 5,000 mm—the highest in Europe—and snow patches persist into June down to altitudes as low as 1,000 m a.s.l. (photos: Philip Hughes).

Glacial and Periglacial Environments

230

0

Kredarica Tr Triglav iglav

00

00

26

24

0 250

2539m

2863m

2700 25

2400

0

Moraines and glacier extent in 1850

230

0

00

230

N

Glacier extent in 1952 Glacier extent in 1999

0

1000 m

Fig. 12.3. Map of the Zeleni Sneg glacier on Mount Triglav, Slovenia, depicting ice retreat since the mid-nineteenth century. Based on sources cited in the text.

The coastal Dinaric Alps are the wettest part of Europe. Here, mean annual precipitation can exceed 5,000 mm and the absolute annual maximum of 8,063 mm was recorded in 1937 (Magaš 2002). Despite such high precipitation values, only a few glaciers exist in locally favourable topographic settings, since the snowline occurs at above the highest peaks of this area (Hughes 2007). On Mount Triglav (2,864 m a.s.l.), in the Julian Alps of Slovenia, the Zeleni Sneg glacier is situated on the northern slopes between c.2,550 and 2,400 m. Figure 12.3 illustrates the fluctuations in the size of this glacier over the last 150 years and highlights its dramatic retreat throughout the second half of the twentieth century (Sifrer 1963). By 1995 the glacier covered an area of only 0.0303 km2 (Gabrovec 1998). Increases in summer temperatures and maximum daily temperatures from May to September between 1954 and 1994 are closely correlated with the retreat of the glacier front and the reduction in ice thickness (Gams,1994). Small glaciers and permanent snowfields are present further south in the Dinaric Alps. A small glacier called Debeli Namet, is present on the north face of Sljeme (2,455 m a.s.l.) in the Durmitor Massif, Montenegro (Figure 12.4) (Hughes 2007). It is possible that this glacier represents the remnants of a Little Ice Age glacier, since the lower margins are bounded by arcuate subrounded boulder ridges and a distinct trimline marks the boundary between freshly scoured bedrock and older glaciated terrain (Figure 12.4). Small glaciers and

357

permanent snowfields also occur in the Prokletije Mountains (2,694 m a.s.l.) on the border of Montenegro and Albania. Several glaciers were mapped in the First World War by the Ministry of Arms of the Austro-Hungarian Empire (Roth von Telegd 1923), although the current state of these features is unknown. If glacier ice is still present in the Prokletije, then these would represent some of the most southerly glaciers in Europe—at a similar latitude to the Calderone glacier in Italy discussed below.

The Italian Appenines The Apennine Mountains span the length of the Italian Peninsula. The highest peak is Corno Grande (2,912 m a.s.l.) in the Gran Sasso of Abruzzi where the Calderone glacier is found. This is Europe’s southernmost glacier. The glacier occurs below the regional snowline in a steep-sided, north-facing cirque (Gellatly et al. 1994) where the local topography creates a microclimate suitable for glacier development. The glacier has retreated through the twentieth century and between 1916 and 1990 its volume is estimated to have been reduced by about 90 per cent and its area by about 68 per cent (ibid.). The future of the Calderone glacier is therefore in doubt and in common with other glaciers of southern Europe, it will soon disappear (J. M. Grove 2004: 218). Glaciers are not present anywhere else in the Italian Apennines. However, permanent ice has been reported in Sicily from the Grotta del Gelo (Cave of Frost) at c.2,030 m a.s.l. on Mount Etna (Marino 1992). The cave is situated well below the regional snowline and again the presence of ice is likely to be controlled by local conditions. The lower limit of discontinuous permafrost is estimated to be at c.2,600 m in the Apennines, based on rock glacier activity and the occurrence of perennial snow patches (Dramis and Kotarba 1994). Several active periglacial features have been described in the Gran Sasso Massif. These include patterned ground, stone-stripes, solifluction lobes, ploughing blocks, and nivation landforms reported above the treeline (c.2,000 m) by Gentileschi (1967a, b) and Kelletat (1969), and down to about 1,600 m by D’Alessandro et al. (2003). The lower zone is characterized by the seasonal occurrence of spring snow avalanches which can destroy any woodland in their path (D’Alessandro and Pecci 2001; D’Alessandro et al. 2003). This phenomenon was also witnessed by the authors in the Pindus Mountains in May 2003 (Figure 12.5).

358

Philip Hughes and Jamie Woodward

Fig. 12.4. The Debeli Namet glacier on the northern slopes of Sljeme (2,455 m a.s.l.) in Montenegro (photo: Philip Hughes).

The Alpes Maritimes The Alpes Maritimes of southern France and western Italy currently support fifteen small glaciers (Federici and Pappalardo 1995). These glaciers are the southernmost of the Alpine chain and some are situated less than 50 km from the Mediterranean Côte d’Azur (Figure 12.1). Thirteen of the glaciers are found in the Argentera Massif, which contains the highest peaks of Alpes Maritimes. The ELA (equilibrium line altitude) of the six largest Argentera glaciers is currently c.2,800 m a.s.l. (Fisinger and Ribolini 2001). In common with other parts of the Mediterranean, all of these glaciers have retreated during the last century and recent glacier behaviour is probably a prelude to the extinction of the glacier ice in the Alpes Maritimes (Federici and Pappalardo 1995; Pappalardo 1999). The lower boundary of discontinuous permafrost in the Argentera Massif occurs at c.2,500 to 2,600 m and active rock glaciers, blockfields, blockstreams, and gelifluction lobes are generally located above these altitudes (Fisinger and Ribolini 2001; Ribolini and Fabre 2006). On the Italian side of the Argentera Massif, seventy-one rock glaciers have been identified. However, most of these are likely to be relict features since some have front elevations as low as 2,350 m—which is well

below the present-day limit of discontinuous permafrost (Ribolini 1999; Fisinger and Ribolini 2001).

The Pyrenees On thirteen peaks in the Pyrenees (Figure 12.6), all over 3,000 m, Serrat (1993) identified forty-one glaciers that covered a total area of 8.10 km2 in 1984. The largest glacier is the Glaciar d’Aneto in the Spanish Pyrenees, which covered an area of 1.32 km2 in 1984. Glacier extent has declined since then and Chueca et al. (2005) reported that the Glaciar d’Aneto had retreated to cover an area of only 0.904 km2 . The glaciers of the Pyrenees are situated close to threshold conditions for glaciation and their development and morphology are strongly influenced by the effects of topography on aspect and shadowing (Chueca and Julián 2004). Most glaciers have suffered negative mass balance in recent years and the Maladeta glacier had a mean specific net balance of −811 mm and −1,102 mm (water equivalent) in 2001/2 and 2002/3, respectively (World Glacier Monitoring Service 2003, 2005). The Pyrenees glaciers appear not to be in equilibrium with modern climate, and further rises in temperature are likely to raise the glacier threshold above the highest cirques.

Glacial and Periglacial Environments

359

Fig. 12.5. The effects of snow avalanching on beech trees on the northern slopes of Mount Tymphi, Greece (photo: Jamie Woodward).

Serrano et al. (1999) identified thirteen active rock glaciers in the central Pyrenees, with five on the northern slopes and eight on the southern slopes. The south-facing rock glaciers had formed on peaks that exceeded 2,950 m and a lower limit of discontinuous permafrost was identified at c.2,800 m a.s.l. on the north-facing slopes. However, in common with the glaciers, conditions for permafrost are becoming increasingly marginal in the high mountains of the Mediterranean. This was highlighted for this region by Chueca and Julián (2005) who undertook a ten-year geodetic survey of the Besiberris rock glacier in the central Pyrenees. They found a marked thinning of a central ice core that they attributed as a response to recent climatic amelioration.

Iberia On the Iberian Peninsula south of the Pyrenees, few glaciers occur today and there is some debate as to whether any exist at all. González-Suárez and Alonso (1994) reported the presence of a small glacier in the

Picos de Europa (2,651 m), the highest Massif of the Cantabrian Mountains, and attributed its existence to high precipitation and shading. However, Frochoso and Castañón (1995) argued that these features represent fossil ice bodies inherited from the LIA and that today, only perennial snow patches and sporadic permafrost occur in this area. On many of the highest Iberian peaks, snow persists for more than 220 days of the year. Snow accumulation is an important influence on present-day geomorphological and biogeographical processes—especially on northern and eastern slopes (Palacios et al. 2003). In the Peñalara Massif (2,428 m a.s.l.) of central Spain, for example, nivation niches, protalus ridges, and nivoalluvial fans are actively forming today. In southern Spain, glacier ice existed until the late twentieth century in the Sierra Nevada at the foot of the north wall of Picacho del Veleta (3,398 m) in the Corral Veleta cirque (García 1996). Drilling investigations have shown that ice still exists within the talus in this cirque down to c.3,050 to 3,100 m (Gómez et al. 2001, 2003). The buried ice in the Corral del Veleta is

360

Philip Hughes and Jamie Woodward

Toulouse

France Ga

ro

nn

e

Biarritz

Bilbao Vignemale

Perpignan Andorra Maladeta

Zaragoza

Spain

Eb

ro

Barcelona

N 0

Glacierized areas

50 km

Fig. 12.6. The distribution of currently glacierized areas in the Pyrenees. All the glaciers are small cirque glaciers and the shaded areas represent general glacierized areas and not the glaciers themselves. Based on sources cited in the text.

significant because it represents the southernmost permafrost remnant in Europe (Gómez et al. 2001) and periglacial processes are likely to be active in other areas of the Sierra Nevada. In fact, the whole summit area of the Sierra Nevada, which culminates at the peak of Mulhacen (3,478 m), represents a marginal area of discontinuous permafrost (Gómez et al. 1999) characterized by miniature patterned ground, terracettes, and active solifluction lobes (Simon et al. 1994; Gómez-Órtiz and Salvador Franch 1998). In the Serra da Estrela of Portugal (Figure 12.1), Brosche (1978) and Daveau (1978) have reported the occurrence of a periglacial belt on the highest slopes, but they present conflicting opinions of its extent. However, Vieira et al. (2003) have argued that periglacial activity related to permafrost processes, sensu stricto French (1996), is not present in the Serra da Estrela, but recognize that frost action and associated cryogenic landforms are active. These include miniature sorted stripes and nets, microforms related to needle-

ice activity, incipient solifluction lobes, terracettes, signs of microgelivation, and evidence of upfreezing of granules (Vieira et al. 2003). Here, as in many other Mediterranean mountain areas, it would appear that, whilst permafrost is absent, cryogenic processes play an active role in sediment production and landform development.

The Atlas Mountains The Atlas Mountains are the most southerly mountains in the Mediterranean basin and extend over 1,000 km from Morocco to Tunisia. The highest mountains occur in the High Atlas in Morocco where Mount Toubkal reaches 4,165 m. Further north in the Middle Atlas the highest peak of Jbel Bou Iblane reaches 3,340 m. Elsewhere, in the Rif, Tell, and Saharan Atlas ranges, peaks do not exceed 2,500 m. Periglacial features such as solifluction lobes, rasentrappen (German: ‘sea of rock’; a mantle of angular shattered boulders),

Glacial and Periglacial Environments

361

Fig. 12.7. An unglaciated periglacial surface covered in felsenmeer (blockfield) on Ouanoukrim (4,063 m a.s.l.) in the Atlas Mountains. Glacier ice occupied the cirque in silhouette in the middle distance. Mount Toubkal (4,165 m a.s.l.), the highest peak in the Atlas Mountains, is in the background (photo: Philip Hughes).

thufurs (mounds formed by heaving of the ground surface), polygons, stone stripes, and felsenmeer are active today above 2,000 m in the High Atlas (Couvreur 1966) (Figure 12.7). Frosts are frequent and winter temperatures can fall below −20◦ C with snow lingering until early summer on the highest peaks (Robinson and Williams 1992). At one site below the northern cliffs of the Tazaghart plateau, in the Toubkal area, perennial snow has been reported by Smith (2004).

Holocene and Little Ice Age Glacial and Periglacial Environments Climate during the Holocene (the last 11,500 years) has been relatively stable compared to the rapid and pronounced shifts in temperature that took place during the Last Glacial stage (cf. Dansgaard et al. 1993;

Chapter 4). However, the Holocene has been characterized by climatic oscillations of sufficient magnitude to cause significant glacier oscillations in various parts of the world including the Mediterranean region. The most recent major glacier advance, recorded in most mountain ranges across the world, occurred during the LIA. The term ‘Little Ice Age’ is widely used to describe a period of glacial advance between c.1550 and 1800 (Lamb 1977: 104) and was characterized by global cooling of the order of 1–2◦ C (J. M. Grove 2004: 591). Glaciers advanced in most glaciated European mountain regions, such as in Iceland, Scandinavia, and the Alps. Glacier advances in the Alps have been linked with extreme weather types in the Mediterranean region during the LIA, including extreme rainfall, heavy snow, and intense cold (A. T. Grove 2001). The impacts on flooding and river basin processes are discussed in Chapter 11. Glaciers are known to have advanced in

362

Philip Hughes and Jamie Woodward

the Pyrenees, Alpes Maritimes, Italian Apennines, and in Turkey during this period. In some areas, such as southern Spain, glaciers that were present during the LIA no longer exist.

Turkey In the Turkish mountains, glaciers appear to have begun retreating in the early twentieth century, a process that has accelerated since the 1930s (Erinç 1952; Güner and Emre 1983; Çiner 2004). However, detailed knowledge of the Holocene history of most Turkish mountain glaciers is not yet available.

The Balkan Peninsula In the Balkans, historical documents record severe cold and heavy snowfalls during the LIA, especially during 1676–1715 and 1780–1830 (Xoplaki et al. 2001). In the winter of 1686/7, almost every region of Greece experienced severe cold and heavy snow and Lake Pamvotis at Ioannina remained frozen for more than three months (Chapter 4). The winter of 1708/9 was also particularly severe in the Balkans and is widely regarded as one of the most severe of the last 500 years, resulting in widespread famine, plague, and death in the former Yugoslavia (Xoplaki et al. 2001). Few studies, however, present evidence for the effects of climatic deterioration during the LIA on glacial and periglacial landform development in the mountains of the Balkans. An exception is at the Zeleni Sneg glacier on Triglav in the Julian Alps, Slovenia. This glacier covered an area of c.40 hectares in the nineteenth century (Sifrer 1963) but, as noted earlier, had retreated to only a few hectares by 1995 (Gabrovec 1998) (Figure 12.3).

The Italian Apennines The Calderone glacier, in the central Apennines, was much more extensive during the LIA compared with today. Moraines at the threshold of the Calderone cirque have been attributed to the LIA and occur 1,000 to 1,050 m above moraines attributed to the Würmian glacial maximum (Federici 1979; Gellatly et al. 1994). More recently, Giraudi (2003, 2004) recognized a complex of Holocene Neoglacial glacial units in this area and identified five phases of glacier advance, using radiocarbon dating to establish the age of interbedded soils. Giraudi (2004) proposed that ice had completely melted in the Calderone cirque by 9,00014 C years BP and subsequent glacier formation and expansion occurred after 3,890 ± 60, 2,650 ± 60, 1450 ± 40, and 670 ± 4014 C years BP. A final glacial advance during the fourteenth century AD covered these earlier units and formed

the Neoglacial glacier maximum, thus mirroring glacier expansion elsewhere in Europe during the Little Ice Age (Giraudi 2003, 2004). Variations in Holocene glacial activity and periglacial processes in the Apennines show that winter temperatures were 3.0◦ C lower than today between c.790 to 150 calendar years BP and 1.2◦ C higher than today between c.5,740 and 5,590 calendar years BP (Giraudi 2005).

The Alpes Maritimes Federici and Stefanini (2001) carried out a lichenometric study on glacial deposits in the high mountains of the Alpes Maritimes to examine the nature of LIA glacier fluctuations. The most extensive glaciers of historical times occurred between the thirteenth and fourteenth centuries (c.AD 1215–50 and AD 1310–50) and moraines were deposited at higher elevations between c.AD 1640–90, 1760–1785, and 1825. All the glacier advances occurred during the LIA and the first correlates with glacier expansion elsewhere in the Alps (cf. J. M. Grove 2004). The mean ELA of the most extensive LIA glaciers in the Argentera area was c.2,650–2,700 m (Fisinger and Ribolini 2001). Earlier Holocene glacier advances are recorded by moraines at slightly lower altitudes and were formed by glaciers with a mean ELA of c.2,600–2,650 m.

The Pyrenees The Pyrenees represent a classic area where glaciers are known to have advanced during the LIA and, in common with the Alps, glacier fluctuations were documented by observers as early as the eighteenth century. In the Hautes-Pyrenees, radiocarbon dating of sediments from a bog at c.2,100 m in the Cirque de Troumouse showed that moraines immediately downvalley formed during the Holocene before c.519014 C years BP (Gellatly et al. 1992). Neoglacial events are also indicated by younger silt-rich sediments in the bog and two younger moraine sets on the cirque floor at altitudes of c.2,250 to 2,350 m. Little Ice Age glaciers were more restricted in extent, and are recorded by moraines at c.2,360 and at c.2,650 m, close to the cirque wall (Gellatly et al. 1992). Observations of LIA glacier characteristics were made by Louis-François Ramond de Carbonnières in the late eighteenth century—when glaciers were significantly more extensive than today (Grove and Gellatly 1995; J. M. Grove 2004) (Figures 12.8 and 12.9). Glacier retreat during the early nineteenth century was followed by a mid-nineteenth-century readvance, albeit smaller than the late eighteenth century readvance. This was followed by a period of major glacier contraction

Glacial and Periglacial Environments

Glacier in 1993

Crest line

Névé

River

Peak

Moraine

363

Approximate ice margin, 1798 Approximate ice margin, 1865 Approximate ice margin, 1927

ête

Cr

Cairn used in 1927 survey (Gaurier 1933a)

ique Longue du P

Grottes (Russell)

Point Chaussenque 3147m

Glacier d’Ossoue

Cr

ête

du

nt Mo

fer

rat

Ga

ve

d’O

sso

ue

Pic Centrale

Glacier de Mont-Ferrat

N

0

1 km

Fig. 12.8. Map of the Glacier d’Ossoue, the largest modern glacier in the Pyrenees, showing former ice margins in 1798, 1865, 1927, and 1993 (modified from an original sketch map compiled from aerial photographs and field mapping in Grove and Gellatly 1995).

between 1864 and 1870 (Michelier 1887). At Glacier du Pays Baché—on the eastern side of the Néouville Massif in the Hautes-Pyrenees—Michelier estimated that the glacier lost c.8,400,000 m3 in volume as a result of reduced precipitation and warmer winters after the middle of the century until the 1880s. The late nineteenth century was characterized by yet another advance and Bonaparte (1890) presented photographic evidence of increased ice thickness between 1880 and 1890. J. M. Grove (2004: 199), too, cites evidence for increased accumulation for the period 1885–95 at Glacier d’Ossoue on Vignemale (Figure 12.8). Whilst glacier recession has dominated the twentieth century, it has not been continuous, as glaciers either stabilized or advanced slightly during the intervals from 1906– 11, 1926–7, 1944–5, 1963–4, and 1978–9 (Grove and Gellatly 1995; J. M. Grove 2004). In the Posets Massif of Spain, the second highest in the Pyrenees, Serrano et al. (2002) identified six phases

of LIA glacier behaviour with the maximum extension of glaciers reached between 1600 and 1750. A minor advance occurred between 1905 and 1920, whilst the interval 1920–80 was generally characterized by glacier recession and a particularly rapid recession between 1980 and 2000 (Serrano et al. 2002). Overall, the total reduction in glacier surface area since the LIA is estimated at 39.2 per cent in the Posets Massif (Chueca et al. 2005). Figure 12.9 shows the extent of LIA glacial deposits presented in J. M. Grove (2004: 190) on the Maladeta Massif, the highest Massif of the Pyrenees, and to the east of the Posets Massif. A recent study by Chueca et al. (2005) reconstructed the LIA evolution of the Maladeta glacier; the largest in this area. This glacier reduced in size by 35.7 per cent from 152.3 hectares in 1820–30 to 54.5 hectares in 2000 and the ELA rose by 255 m over the same period. However, glacier retreat was interrupted by at least three phases of

364

Philip Hughes and Jamie Woodward Ice

L. de Paderna

Little Ice Age moraine Lake Watershed

Alba

Maladeta MALADETA

Creguena Aneto

Barrancs Coronas

ANETO Tempestades Tempestades Llosas

N

Selencas

1 km

Fig. 12.9. Map of the glaciers of the Maladeta Massif where some of the largest modern glaciers in the Pyrenees are found (modified from J. M. Grove 2004: 190).

significant stability during the intervals 1820–30 to 1857; 1914–20 to 1934–5; and 1957–81. Ice depletion here was matched by ice loss across the Pyrenees and the rising ELA is characteristic of trends observed in other Mediterranean glaciers.

Iberia The Corral Veleta glacier in the Sierra Nevada is the most celebrated LIA glacier in the Mediterranean (Messerli 1967; García 1996). This glacier reached 800 m in length during the LIA, but never advanced beyond the confines of its cirque and began its retreat in the mid1800s. Further north, on the northern slopes of Pico Almanzor (2,592 m), in the Sierra de Gredos of central Spain, Sancho et al. (2001) used lichenometry to show that a protalus rampart, situated at the base of a nivation patch, formed during the LIA. This evidence supports other studies that suggest that there was a substantial increase in snow cover in the Sierra de Gredos during the LIA (e.g. Toro et al. 1993).

Pleistocene Glacial and Periglacial Environments As Figure 12.10 shows, many Mediterranean mountains were glaciated during the Pleistocene (Messerli 1967; Hughes et al. 2006a; Hughes and Woodward, 2008). In some areas, this was restricted to the highest cirques, whilst in others ice fields formed and valley glaciers extended many kilometres into lowland areas. Significantly, the periglacial zone was lowered considerably during Pleistocene cold stages and, in conjunction with glacial and fluvial processes, it helped to shape much of the upland Mediterranean landscape we see today. Many Mediterranean mountain landscapes can therefore be viewed as products of Pleistocene glaciation and periglacial activity in addition to the influence of tectonic, fluvial, and associated processes. In several areas, glacial and periglacial processes have interacted with karstic processes to form some of the best examples of glacio-karst landscapes (Chapter 10). Notable examples occur in the Pindus Mountains, Greece (Waltham 1978; Woodward et al. 2004; Hughes et al. 2006b) and in the Picos de Europa of north-west Spain (Smart 1986).

Turkey and the Near East In the Near East, evidence of former glaciation was first noted in the late nineteenth century by Diener (1886) in the mountains of Lebanon. According to Messerli (1967), glacial evidence is particularly well preserved on Qornetes Saouda (3,088 m), the highest peak of the Jbel Liban, and on Mount Hermon (2,814 m) in the southern Anti-Lebanon. Further north in Turkey, evidence of Pleistocene glaciation can be found throughout the main mountain ranges. On Mount Ararat, in the eastern Taurus Mountains, the snowline was depressed to c.3,000 m a.s.l. during the Pleistocene and the volcanic cone was covered by an ice cap of about 100 km2 (Blumenthal 1958). Wright (1962) estimated that Pleistocene snowlines across this part of Turkey were depressed by between 1,200 to 1,800 m and suggested that much of this depression can be attributed to increased snowfall (relative to today), in addition to lower summer temperatures. The Taurus and Pontic mountains of Turkey were extensively glaciated during the Pleistocene. U-shaped valleys, moraines, roches moutonnées, and glacial lakes are abundant throughout this region (Çiner et al. 1999; Çiner 2004; Akçar and Schlüchter 2005). In the Aladag Massif of the central Taurus, an ice cap on the

Glacial and Periglacial Environments

Glaciated areas

0

365

1000 km

Fig. 12.10. Map of the former extent of Pleistocene glacial and nivation features in the Mediterranean (after Messerli 1967).

Yedigöller plateau (c.3,200 m a.s.l.) covered 40 km2 and outlet glaciers flowed into valleys to the north-east, west, and east (Bayari et al. 2003). Here there is evidence of seven episodes of glacial advance and retreat, although just two separate suites of moraines have been identified based on contrasting moraine ‘freshness’, with the best preserved above 1,800 m, and others extending down to 1,200 m (ibid.). Relict rock glaciers have also been widely preserved and in the central Taurus two generations of rock glaciers are concentrated between 2,050 and 2,500 m a.s.l. (Arpat and Özgül 1972). A very recent development in Turkey has seen the application of cosmogenic exposure dating to Pleistocene glacial records in the Kaçkar Mountains (Akçar et al. 2007, 2008) and in the western Taurus Mountains (Sarıkaya et al. 2008). These studies are providing important new insights into long-term glacier behaviour in the eastern Mediterranean and they have raised important new questions about the palaeoclimatic conditions associated with glacier development in this region (Hughes and Woodward 2008).

The Balkans Glaciers formed in the mountains of Greece, Albania, and the former Yugoslavian states on multiple occasions during the Pleistocene (Hughes et al. 2006a).

In Greece, glacial deposits and landforms have been recorded by numerous workers (Niculescu 1915; Sestini 1933; Mistardis 1952; Messerli 1967; Pechoux 1970; Fabre and Maire 1983; Palmentola et al. 1990; Boenzi et al. 1992; Smith et al. 1997), and in many areas moraines and limestone pavements are very well preserved (Figures 12.11 and 12.12). However, the timing of glaciation in Greece has only recently been established. On Mount Tymphi, in the northern Pindus, Woodward et al. (2004) dated calcite cements in tills using U-series methods and found that the oldest, most extensive glacial deposits were older than 350,000 cal. years BP (Figure 12.13). Subsequent work by Hughes et al. (2006b) has identified evidence for three glacial phases on Mount Tymphi by combining detailed geomorphological mapping with an extended programme of U-series dating. Hughes et al. (2006b) correlated the various glacial and periglacial units recorded on Mount Tymphi with cold stage intervals recorded in the pollen stratigraphy from Ioannina about 40 km to the south (Tzedakis 1994; Tzedakis et al. 2002; Chapter 4). The Ioannina sequence was then used as a parastratotype to define a glacial chronostratigraphy for the Pindus Mountains and to facilitate correlations with the marine isotope record (Table 12.2). The extent of the former glaciers during the different glacial stages on Mount Tymphi is illustrated in Figure 12.15. A key outcome of this work is

366

Philip Hughes and Jamie Woodward

Fig. 12.11. Moraines at c.1,700 m a.s.l. in the Vourtapa valley above the village of Skamnelli on Mount Tymphi, Greece (photo: Philip Hughes).

that the glaciers of the last cold stage in Greece were much smaller than during earlier Middle Pleistocene glaciations. Well-developed periglacial forms are often present inside former Pleistocene glacier limits. For example, thick accumulations of talus (Figure 12.14) inside of the Vlasian Stage (MIS 6) glacier limits (Figure 12.15, Table 12.2), are the product of periglacial weathering during the last cold (Tymphian) stage and, to a lesser extent, through the Holocene. Hughes et al. (2006c) found evidence of a similar glacial sequence to Mount Tymphi on neighbouring Mount Smolikas (2,637 m). Here, however, there is also evidence in the highest cirques for a fourth and later glacial phase when small cirque glaciers developed. These small glaciers on Smolikas had an ELA of c.2,420 m and covered a total area of <0.5 km2 . This phase of glaciation is likely to have occurred after the glacial maximum of the Tymphian Stage (Table 12.2) and possibly during the Younger Dryas (11,000–10,000 14 C years BP) (ibid.).

The glacial history of Mount Olympus (2,917 m a.s.l.), the highest mountain in Greece, has been documented by Smith et al. (1997). The Olympus glaciers extended to the piedmont zone during the most extensive glacial phase down to altitudes as low as 500m a.s.l. (Figure 12.16). Smith et al. (1997) proposed a tentative chronology for glaciation on Mount Olympus by correlating soils on glacial deposits with dated soils in the river deposits of the Larissa basin. The oldest and most extensive glaciation was correlated with soils older than 200,000 cal. years BP, leading Smith et al. (1997) to suggest that this glaciation occurred during MIS 8. A second phase of glaciation, characterized by upland ice and valley glaciers that did not reach the piedmont, was correlated with MIS 6. During the last major glacial phase, glaciers were restricted to valley heads, and this glacial phase was correlated with MIS 4 to 2. Smith et al. (1997) suggested that a further set of moraines at c.2,200 m in the high cirque of Megali Kazania may be Holocene Neoglacial features. However, (Hughes et al. 2006c) have argued that features at a similar altitude on Mount

Glacial and Periglacial Environments

367

Fig. 12.12. Limestone pavement on Mount Tymphi, Greece (photo: Philip Hughes).

Smolikas (2,632 m a.s.l.) in the Pindus Mountains were Late Glacial in age. The chronological model suggested for Olympus by Smith et al. (1997) differs from that established for the Pindus Mountains to the west. However, the morphostratigraphical sequences are very similar and, given the short distance between these areas, it is likely that the Olympus and Pindus sequences correlate. The dating framework proposed by Woodward et al. (2004) and Hughes et al. (2006b) for Mount Tymphi is likely therefore to provide a good approximation for the age of the glacial sequences on Mount Olympus, but this can only be confirmed by further work. Almagià (1918) provided one of the earliest accounts of the evidence of former glaciation in the mountains of central and northern Albania. The largest glaciers formed in the Prokletije Mountains of northern Albania and early work by Roth von Telegd (1923) estimated that Pleistocene valley glaciers extended for some 35 km over the border into Montenegro near the towns of Gusinje and Plav. In fact, the town of Plav is built on moraines that dam a large lake at c.900 m a.s.l. and the age of these moraines is currently the focus of research by the authors. In southern Albania, evidence for glaciation was reported by Louis (1926) from the mountains of Nëmerçka (2,495 m), on the

Epirus border, and Mali i Lunxheriës (2,200 m), south-west of Gjirokastër. More recently, Palmentola et al. (1995) have reported the presence of relict rock glaciers above 1,700 m that were set within more extensive glacial features in the Prokletije Mountains. The geochronology of glaciation in this area is yet to be established. In the former republics of Yugoslavia, Jovan Cviji´c was a pioneer of glacial research and produced remarkably detailed analyses of the glacial history of the region (e.g. Cviji´c 1900, 1917). More recently, Menkovic et al. (2004) have compiled evidence of glaciation in the mountains of Serbia, Macedonia, and Montenegro. Glacio-karst landscapes dominate many of the highest uplands and glacial landforms are often well preserved (Figures 12.17 and 12.18). Some of the largest Pleistocene glaciers formed in the Prokletije Mountains, where glaciers extended over 35 km, forming U-shaped valleys, moraines, and numerous glacial lakes (Cviji´c 1913). Further west, on Mount Durmitor (2,523 m a.s.l.), in Montenegro, Alpine-type valley glaciers descended from the highest peaks into a large plateau ice field above canyons incised by the Piva and Tara rivers (Marovic and Markovic 1972; Nicod 1968; Menkovic et al. 2004). A major ice cap also formed on Maganik to the south and, here, the glacial limits on

368

Philip Hughes and Jamie Woodward

Fig. 12.13. Cemented till on Mount Tymphi, Greece, that has yielded ages >350,000 years BP using U-series (Woodward et al. 2004, Hughes et al. 2006b) (photo: Philip Hughes).

the western slopes were first traced by Liedtke (1962). Near the Adriatic coast on Mount Orjen (1,895 m a.s.l.), glacial cirques and valleys exploited older karstic forms to form extensive glaciers (Penck 1900; Sawicki 1911; Menkovic et al. 2004). The extensive glaciation

on Mount Orjen (1,895 m) was probably a function of very high precipitation and today modern values here exceed 5,000 mm per year (Magaš 2002). Marjanac and Marjanac (2004) reviewed the evidence for glaciation in the coastal Dinaric Alps of

Glacial and Periglacial Environments

369

TABLE 12.2. Correlation table showing the relationship between the fragmentary glacial sequence in the Pindus Mountains, Greece, and the continuous lacustrine parasequence in the nearby Ioannina 249 and 284 cores (Chapter 4) Age (× 1,000 years) 11.573.9–11.5

MIS

Ioannina Parastratotype (IN 249/284) boundary (IN 249)

73.9–83.0∗ 88.5∗ –83.0∗ 104.5∗ –88.5∗ 111.0∗ –104.5∗

1 2 3 4 5a 5b 5c 5d

Interstadial 2 Stadial 2 Interstadial 1 Stadial 1

45.88 m

126.6∗ –111.0∗

5e

Metsovon

59.00 m

189.9–126.6∗

6

244.2–189.9

7

303–244.2

8

339–303

Holocene

Pindus Chronostratigraphy

Local Stratotype

17.25 m

76.00 m

Tymphian Stage

Tsouka Rossa Member 39◦ 58 45 N, 20◦ 50 40 E, 2025 m a.s.l.

Vlasian Stage

Vourtapa Member 39◦ 55 50 N, 20◦ 51 10 E, 1650 m a.s.l.

IN-26 Zitsa IN-23a ?

362–339

9a–e Katra IN-17 Pamvotis 10

423–362

11

478–423

12

Dodoni I, II

No record of glaciation during this interval has been found in the Pindus Mountains. However, Vlasian Stage glaciers are likely to have overridden any glacial deposits formed during MIS 10 and 8.

? 162.75 m 184.00 m

Skamnellian Stage

Kato Radza Member 39◦ 54 08 N, 20◦ 50 40 E, 984 m a.s.l.

Note: MIS = Marine Isotope Stage. Names in the third column based on Tzedakis et al. (2002)—all other names for the Ioannina sequence are from Tzedakis (1994). ∗ Interval dates from Tzedakis et al. (2002)—all other dates from orbitally tuned marine isotope records (Imbrie et al. 1984, Martinson et al. 1987) (from Hughes et al. 2005). Sources: Tzedakis (1994); Tzedakis et al. (2002).

Croatia. They describe features on some of the Croatian coast and islands in the Adriatic, which they argue are glacial in origin. These include kame-terraces on the Krk and Pag islands to the west of the Velebit Mountains, as well as glacial and periglacial deposits on the mainland coast nearby at Novigradsko More and Karinsko More. They (ibid.) attribute the coastal glacial deposits to a glaciation during the Early or Middle Pleistocene, but have acknowledged that more work is needed to clarify the chronology of the Croatian glacial sequences. Whatever the age of the glacial deposits, their presence at such low altitudes in the Mediterranean is of major significance.

The Italian Appenines The most comprehensive geochronological framework for the Pleistocene glacial sequence in the Italian

Appenines is in the Gran Sasso. Here, based on the radiocarbon dating of ice-dammed lacustrine sediments, Giraudi and Frezzotti (1997) have demonstrated that the maximum glacier extent of the last cold stage (Würmian) occurred prior to 22,680 ± 630 14 C years BP. At the glacier maximum, ice in the Campo Imperatore area of the Gran Sasso extended 10.5 km down-valley and covered an area of 19 km2 with an ELA of c.1,750 m. A series of recessional moraines and rock glaciers in the Gran Sasso are thought to correspond to periods of glacier stabilization or readvance between 20,000 and 10,000 14 C years BP. Based on rock glacier evidence, Giraudi and Frezzotti (1997) concluded that mean annual temperatures during the Last Glacial stage were 7.3 to 8.3◦ C lower than today. They extrapolated this temperature depression to the ELA of the contemporaneous valley glacier in the Campo Imperatore area and, based on the well-established relationship between accumulation and

370

Philip Hughes and Jamie Woodward

Fig. 12.14. Well-developed screes within the limits of Vlasian Stage glaciers (Figure 12.15, Table 12.2) on the southern slopes of Mount Tymphi, Greece (photo: Jamie Woodward).

precipitation at the ELA of modern glaciers (cf. Ohmura et al. 1992), they concluded that snowfall was similar to today. In this area Middle Pleistocene glacial deposits are usually less well preserved and are often strongly eroded with a much less pronounced morphology. Kotarba et al. (2001) presented U-series ages for calcite cements within moraines of 135,000 ± 10,000 cal. years BP. Thus, given that the cements formed after the deposition of the host glacial deposits, the moraines were correlated with the Rissian Stage of the Alps (Kotarba et al. 2001). At least two major glacial advances are recorded in the Gran Sasso area, with readvance also recorded during the Würmian Late Glacial. Glacial deposits that extend lower than moraines assigned to the Würmian Stage have been noted in many other areas of the Italian Apennines but they are commonly fragmentary and their age has not been established (Federici 1980). As in the Campo Imperatore area, they are often partially cemented and show evidence of prolonged weathering.

Corsica Glacial features in the mountains of Corsica were first reported by Pumpelly (1859) and the most recent published studies include those by Heybrock (1954), Letsch (1956), and Conchon (1978, 1986). In contrast to many of the glaciated mountains in the Mediterranean, those in Corsica are formed in granite, and because of the resistant nature of this lithology, the cirques, arêtes, and roches moutonées are particularly pronounced and well preserved. Evidence obtained from sediment cores, collected up-valley of the most recent moraines, has shown that the highest cirques contained small glaciers between 15 and 14,000 14 C years BP, but have been free of ice since the Allerød Interstadial at c.12,500 14 C years BP (Conchon 1986). Kuhlemann et al. (2005) have used geomorphological evidence to reconstruct the Pleistocene glaciers of Corsica. Large ice fields and valley glaciers formed during the Würmian glacial maximum and some glaciers were

Glacial and Periglacial Environments

Fig. 12.15. The extent of Middle and Late Pleistocene glaciers on Mount Tymphi, Greece.

371

372

Philip Hughes and Jamie Woodward

A

KATERINI

Mav

rone

ri Riv

R M

ro av

ne

iv

er

er

ri

End moraines (lateral/terminal, recessional)

Xerolakki

Areas of ice disintegration (stagnation, collapse) Outwash and alluvium 0

N

Geomorphic contact

2 km

Xerola

kki

Boundary fault

B Unit 3 20

00

0

20km

Thessaloníki

m

Unit 3

M

a

o vr

lo

ng

us

A Katerini

Unit 3

Aegean

Olympus Olympus 2911m 2911 m

Unit 2/Unit 1

B

vra

tza

200

0m Undissected upland (stage 1)

Ma N

Sea

0

2 km

Cirque Glacial deposits

Lárisa

Fig. 12.16. Glacial geomorphological maps of Mount Olympus, north-eastern Greece (modified from Smith et al. 1997).

Glacial and Periglacial Environments

373

Fig. 12.17. Moraines at c.1,000 m a.s.l. in Duboki Do, above the village of Ubli on Mount Orjen, Montenegro (photo: Philip Hughes).

up to 14 km long. The ELA of the Würmian glaciers was between 1,400 and 1,750 m, with variations attributed to precipitation differences. Mean annual temperatures during the Würmian glacial maximum were 8◦ C lower compared with modern values. This finding is similar to recent data from Greece, but good dating control has not been firmly established for the bulk of the Coriscan glacial record. Cosmogenic isotope exposure dating offers much potential as preliminary investigations have shown that the Corsican granites are suitable for this method (Hewitt 2002).

The Alpes Maritimes During the Würmian Stage, the Alpes Maritimes were covered by ice that was contiguous with the main Alpine ice sheet—extending over an area of c.126,000 km2 (Ehlers 1996). The glacial geomorphology of the Argentera area has been described by Ribolini (1996). The valleys of this area are deeply scoured and large moraines were deposited in their lower parts. The Würmian glacial maximum in the Alps occurred between c.28,000 and 20,000 14 C years BP in the northern Alps and at c.20,000 to 18,000 14 C years BP in the southern Alps (Florineth and Schlüchter 2000). Lacustrine sediments at Lac Long Inférieur, a glaciated cirque

(2,090 m a.s.l.) in the Alpes Maritimes, indicate that the ice had melted by 14,190 ± 13014 C years BP (Ponel et al. 2001). However, in the Argentera Massif, Fisinger and Ribolini (2001) document evidence for several glacial advances during the Late Glacial Substage. Granger et al. (2006) present 10 Be cosmogenic ages of 16.3 ka and 18.8 ka respectively for two moraines in the Gesso Valley and argue that they were part of the same glacial phase. The application of exposure age dating at higher elevations in the Gesso Valley has led to the recognition of moraines of the Younger Dryas age (Federici et al. 2008).

The Pyrenees The Pyrenees were extensively glaciated during the Pleistocene, a fact recognized in pioneering research in the early nineteenth century by Albrecht Penck (1885). These glaciations have been attributed to at least two Pleistocene cold stages—the oldest deposits are correlated to the Rissian Stage and a second, higher, suite of deposits to the Würmian Stage (Barrère 1963). The largest and most powerful Pleistocene glaciers formed on the northern slopes of the Pyrenees in

374

Philip Hughes and Jamie Woodward

Fig. 12.18. Glacial arête between the peaks of Sedlena Greda (2,227 m a.s.l.) and Ranisava (2,081 m a.s.l.) in the Durmitor mountain area, Montenegro (looking south-east from the summit of Sedlena Greda). This ridge separated two major Pleistocene ice flows from the central Durmitor area (photo: Philip Hughes).

France (Calvet 2004). In the Ariège valley, for example, glaciers extended 65 km to an altitude of 370 m a.s.l. (Hérail et al. 1986). The chronology of the last (Würmian) glaciation in the French Pyrenees is based on sedimentological and palynological studies and radiocarbon dating of lacustrine sediments near former glacier margins. This approach has indicated that, during the last glaciation, the glacial maximum in the French Pyrenees was before 38,000 14 C years BP(Hérail et al. 1986; Jalut et al. 1992). This is also the case in the Vosges Mountains of Alsace (Seret et al. 1990). In the Spanish Pyrenees, most glacial deposits are thought to have formed during the last glaciation,

although some isolated glacial deposits have been attributed to earlier periods (Calvet 2004). However, in common with French Pyrenees, the geochronology of the glacial sequences in the Spanish Pyrenees is poorly defined. Correlations have been made on the basis of morphostratigraphical comparisons, and, even delimiting the maximum extent of the last glaciation remains one of the most significant problems of Pyrenean Quaternary geology (García-Ruiz et al. 2003). Nevertheless, it is significant to note that there is evidence that the maximum extent of ice during the Last Glacial stage occurred significantly earlier than for the major ice sheets of Britain and Scandinavia, which reached their

Glacial and Periglacial Environments

maximal extents between 21,000 and 18,000 14 C years BP (Sibrava et al. 1986). Data from various sources suggest that the maximum extent of glaciation during the Last Glacial phase occurred before 30,000 14 C years BP (García-Ruiz et al. 2003), similar to findings in the French Pyrenees to the north. However, recent 10 Be dates from glacial boulders and rock surfaces conflict with the idea of an early glacier maximum. Pallàs et al. (2007) found that the 10 Be dates suggest that the major phase of moraine building in the south-central Pyrenees occurred after 25,000 cal. years BP. Further work using multiple dating techniques is needed to clarify this issue. Deglaciation in the Pyrenees was staggered by a phase of upper valley glaciation between 16,000 and 15,000 14 C years BP and then by a phase of cirque glaciation between 14,000 and 13,000 14 C years BP (Bordonau 1992). The last phase of Pleistocene glaciation is represented by moraines and rock glaciers close to the cirque backwalls and may date from the Younger Dryas between 11,000 and 10,000 14 C years BP (Serrat 1979), in accord with the Alpes Maritimes and the Italian Apennines. In many valleys of the Pyrenees, Würmian Stage glaciers appear to have removed much of the older glacial deposits. However, in some locations, traces of earlier and more extensive glacial deposits exist and these may relate to the Rissian Stage (MIS 10–6) (Barrère 1963). In the eastern Pyrenees, strongly weathered tills are present at lower elevations than those ascribed to the Rissian Stage and are considered to have formed during the early Middle Pleistocene (Calvet 2004). However, the geochronology of the pre-Würmian record is yet to be defined and this is a key issue across much of the Mediterranean region.

Iberia Pleistocene glaciation took place in many of the high mountain areas of Iberia and these are reviewed in Hughes et al. (2006a). Most of the glacial sequences have not been dated, and, given the large number of glaciated sites, only those studies where progress has been made in understanding the geochronology of the glacial record are discussed below. Even though many of the glaciated basins on the Iberian Peninsula do not drain to the Mediterranean Sea, they are located in the same latitudinal belt as the other glaciated terrains discussed in this review and they provide a valuable source of comparison. Some geochronological control is available to the west of the Picos de Europa of northern Spain, in the Redes

375

Natural Park. Here, the most extensive glacial phase was characterized by an ice field with outlet glaciers extending up to 5 km in length. These glaciers descended to c.950 m with snowlines at c.1,550 m. Radiocarbon dating suggests that the maximum phase of glaciation during the Last Glacial stage in this area occurred prior to c.40,000 14 C years BP (Jiménez-Sánchez and Farias 2002). Again, this is well before the global Last Glacial Maximum (LGM), which occurred around 18,000 14 C years BP (CLIMAP Project Members 1976). Glacial landforms have been dated using 36 Cl-based cosmogenic exposure dating in the Sierra de Gredos and Sierra de Guadarrama in central Spain (Palacios et al. 2007). These ages put the local glacier maximum around 21 ka, about the time of the global LGM of MIS 2. Other dated glacial sequences in Iberia include those in the Serra de Queira and Serra de Gêrez in Galicia and northern Portugal (Figure 12.1). Here, Fernadez Mosquera et al. (2000) applied 21 Ne cosmogenic dating to glacially polished surfaces and push-moraine boulders and three glacial phases have been identified. The oldest was dated to before c.238,000 cal. years BP, an intermediate phase to c.130,000 cal. years BP, and the youngest to c.15,000 cal. years BP. This sequence of three major Pleistocene glaciations is comparable to the geochronologies established in Italy and Greece (Giraudi and Frezzotti 1997; Kotarba et al. 2001; Woodward et al. 2004; Hughes et al. 2006b) although the oldest glacial phase appears to be younger than that recorded in Greece and may correspond to a glaciation during the early Rissian (MIS 8). Further south, in the Serra da Estrela of Portugal (1,991 m) (Figure 12.1), glaciation was characterized by the development of a plateau ice cap which fed diffluent glaciers, the longest of which was 13 km and attributed to the Würmian Stage by Daveau (1971). Thermoluminescence ages from fluvioglacial units of between 16.6 ± 2.5 and 10.6 ± 1.6 ka suggest glacial activity in this area during the Late Glacial (Vieira et al.2001).

The Atlas Mountains Quaternary glacial and periglacial features are present throughout the Atlas Mountains of north-west Africa. However, little is yet known of the timing and extent of glaciation in the area (cf. Hughes et al. 2004). Glacial features including cirques, troughs, roches moutonées, riegels, and moraines have been reported from the High Atlas and Pleistocene snowlines have been placed between 3,700 and 3,300 m (Heybrock 1953; Awad 1963; Hughes et al. 2004). The highest peaks in the

376

Philip Hughes and Jamie Woodward TABLE 12.3. Current understanding of the geochronology of Pleistocene glacial deposits in the Mediterranean region Marine Isotope Stage

Alpine/ Northern Europe Chronostratigraphy

Region and approximate age of glacial units

2

Würmian/ Weichselian

Italy: >22,00014 C years BP1 (radiocarbon) Pyrenees: >38,00014 C years BP2 (radiocarbon) Iberia: >29,00014 C years BP3 (radiocarbon); >15,000 years4 (21 Ne cosmogenic) 21,000 years5 (36 Cl cosmogenic) Alpes Maritimes: 19,000 years6 (10 Be cosmogenic) Turkey: 26,000 years7 (10 Be cosmogenic); 20,000 years8 (36 Cl cosmogenic)

6

Late Rissian/ Saalian

Greece: >120,000 years9 (U-series) Italy: >130,000 years10 (U-series) Iberia: >130,000 years4 (21 Ne cosmogenic)

8

Early Rissian/ Saalian

Iberia: >230,000 years4 (21 Ne cosmogenic)

12

Mindelian/ Elsterian

Greece: >350,000 years9 (U-series)

1

Giraudi and Frezzotti (1997); Jalut et al. (1992). 2 García-Ruiz et al. 2003. 3 Jiménez-Sánchez and Farias (2002). 4 Fernandez Mosquera et al. (2000). 5 Palacios et al. (2007). 6 Granger et al. (2006). 7 Akçar et al. (2007), Akçar et al. (2008). 8 Sarıkaya et al. (2008). 9 Hughes et al. (2004); Hughes et al. (2006d); Woodward et al. (2004). 10 Kotarba et al. (2001).

Middle Atlas also show evidence of former glaciation and the regional snowline is estimated at c.2,800 m a.s.l. during the most extensive glacial phase (Raynal et al. 1956; Awad 1963). Periglacial features are also present. Stone polygons, solifluction features, and rock glaciers have been described on Bou Iblane and Jbel BouNaceur, in the Middle Atlas (Raynal 1952; Dresch and Raynal 1953; Awad 1963). Glacial features have also been noted in the Djurdjura Massif (2,308 m) of the Algerian Tell (Figure 12.1) where Barbier and Cailleux (1950) identified cirques, Ushaped valleys, and terminal moraines. To the southeast, in the Aurès Massif, Ballais (1983) noted the presence of moraines above 1,600 m on Jbel Ahmar Khaddou (2,017 m a.s.l.) and Jbel Mahmel (2,321 m a.s.l.). However, the chronology of glaciation in these areas and elsewhere in the Atlas Mountains has not been established and this remains the biggest obstacle to an improved understanding of the glacial history of northwest Africa.

Pleistocene Overview It is clear that very substantial ice masses formed in many Mediterranean mountain areas during Pleistocene cold stages. The glacial deposits and landforms they produced represent important archives of environmental change. However, until quite recently good dating frameworks (Table 12.3) and detailed stratigraphical frameworks had not been established in many key areas. A clear pattern is now emerging whereby the

oldest and most extensive glacial deposits and landforms date from the Middle Pleistocene. In fact, in some areas, at least two phases of Middle Pleistocene glaciation have been identified during intervals equivalent to the Saalian and Elsterian Stages of northern Europe (e.g. Fernadez Mosquera et al. 2000; Woodward et al. 2004; Hughes et al. 2006a, b). Where sound stratigraphical frameworks supported by a robust geochronology do exist for the last cold stage, it has become apparent that glacier maxima in many parts of the Mediterranean mountains preceded the global LGM of MIS 2 by more than 10,000 years (Hughes and Woodward 2008). The small mountain glaciers of the Mediterranean would have advanced and decayed rapidly in response to mass balance fluctuations and they reached their maximum during the last cold stage well before the large ice sheets that covered the Alps and northern Europe. Increased aridity in southern Europe, caused by a strengthening of high pressure systems over the expanding Alpine and north European ice sheets, would have forced glacier retreat in the Mediterranean mountains. The situation is also likely to have been complicated by millennial-scale climate changes recorded in the Greenland ice sheet (Dansgaard et al. 1993) and mirrored in long lacustrine pollen sequences in Italy and Greece (Allen et al. 1999; Tzedakis et al. 2004; Chapter 4). This has been highlighted for the mountain glaciers of Greece by Hughes et al. (2006d) who recognized the potential for multiple phases of glacier advance and retreat during the Last Glacial cycle (Figure 12.19). This analysis identified ten periods, between 115 and

Glacial and Periglacial Environments

377

Fig. 12.19. Climatically favourable conditions for glacier formation based on a combination of pollen and sea surface temperature data. Summary pollen percentage curves from the Ioannina I-284 sequence in north-west Greece, spanning the Last Glacial cycle. Potential intervals suitable for glacier formation are indicated by letters A: major stadials characterized by low arboreal pollen, both including and excluding Pinus and Juniperus; B: intermediate phases between the apices of major stadials and interstadials; and C: intervals characterized by large differences between total arboreal pollen frequencies and arboreal pollen frequencies, excluding Pinus and Juniperus. All other intervals represent major interstadials or interglacials. Both B and C types—the more favourable conditions for glaciation—are indicated by shading. The upper graph depicts variations in the percentage of Neogloboquadrina pachyderma (sinistral) and alkenone-derived seasurface temperatures in marine core MD95-2043 from the Alboran Sea, in the western Mediterranean (from Hughes et al. (2006d).

10 ka, when the climate would have favoured glacier development. The timing of the glacier maxima during earlier glaciations is unclear at present, although large glaciers in the Mediterranean may have been less responsive to rapid climate changes in comparison to those that existed during the last cold stage. Recently published data from Greece show that the transition from glacial to non-glacial conditions took place very rapidly at the end of the last cold stage (Woodward et al. 2008).

Glacial and Periglacial Interactions with Other Geomorphological Systems Glacial and periglacial systems may exert considerable influence on other environments, especially downstream fluvial systems as has been shown for Mount Tymphi, in the Pindus Mountains, where Pleistocene glaciation was a major influence on the longterm behaviour of the Voidomatis River (Chapter 11).

Fluvial sediments transported during cold stages were dominated by materials from the glaciated upland terrains (Bailey et al. 1990; Lewin et al. 1991; Woodward et al. 1992, 1995; Hamlin et al. 2000). The interaction of glacial and fluvial systems has also been explored in the Pineta basin of the central Spanish Pyrenees by Jones (2000). Glaciers and their meltwaters have enhanced karstic processes in many upland areas. In proglacial areas, dolines may be found concentrated in clusters outside terminal moraines (Waltham 1978). Where glaciers occupy only the highest valley and cirque areas, meltwaters often discharge underground before emerging as springs at lower elevations. This process is evident in front of modern and former glaciers in the Pyrenees and was termed ‘Pyrenean type’ glacio-karst by Ford and Williams (1989). This process has also been observed at the Zeleni Sneg glacier, on Triglav in Slovenia, where meltwater disappears through a 280-m deep pothole and resurfaces 1.25 km down-valley (Gams 2001). The deepest potholes in the world have formed in glaciated mountain areas such as the Pyrenees, Alps,

378

Philip Hughes and Jamie Woodward

and Caucasus (ibid.; Chapter 11). Karstic hollows, such as dolines, often promote snow accumulation and can contribute to glacier development. The interaction of glacial and karstic processes is an important and distinctive component of the physical geography of the Mediterranean mountains. Paraglacial effects have also been reported in the region. For example, Lebourg et al. (2003) discussed the importance of high mountain landslides and their sliding mechanisms using a case study of the glaciated Aspe Valley of the Pyrenees. They noted that slope failures occur most frequently where glacial deposits remain and generally at the junction between the till and the underlying strata (Chapter 6). Unstable landslide-prone slopes are likely to be a legacy of glaciation in many Mediterranean mountains, especially where glaciers have retreated during the Holocene and in areas of recent permafrost melting. Over longer timescales, Woodward et al. (2008) have argued that the major Middle Pleistocene glaciations in the Pindus Mountains were responsible for delivering such large volumes of sediment to the fluvial system that reworking of these materials was a major control on river behaviour throughout the Late Pleistocene and Holocene. Finally, periglacial processes also play an important role in a range of geomorphological and sedimentary systems, particularly in terms of debris supply. Frost action can deliver significant quantities of sediment to talus slopes and fluvial systems and frost shattering of rockshelter walls can be an important mechanism of coarse debris accumulation during glacial periods (Woodward 1997; Karkanas 2001 Chapter 6). Thus, even if all the remaining Mediterranean glaciers and areas of permafrost disappear in the next few decades, their legacy will remain, not just as a relict geomorphological record, but also in shaping the sediment supply dynamics in colluvial and fluvial systems.

it is possible that many glaciers will disappear in the twenty-first century. Unfortunately, in many areas, such as in Montenegro and Albania, few detailed studies have been made of recent glacier activity and further work is necessary to understand fully recent glacier dynamics across the Mediterranean (Hughes 2007). Permafrost is present in only the highest Mediterranean mountains and is especially susceptible to minor climatic variations (Gómez et al. 2001). Sporadic and discontinuous permafrost has been identified in the Sierra Nevada and Picos de Europa of Spain, the Pyrenees, the Alpes Maritimes, the Italian Apennines, the Julian Alps, and the Pontic and Taurus Mountains and periglacial processes such as nivation are present in many more mountains. These periglacial environments are also likely to diminish in the Mediterranean mountains during the twenty-first century. It is also clear that climate change will have major implications for high-mountain ecosystems in the Mediterranean region, where plants and animals adapted to cold, alpine conditions now face higher temperatures and a surge of predators and competitors (Krajick 2004; Chapter 23). Whilst contemporary glacial and periglacial environments in most Mediterranean mountains are becoming increasingly rare, the legacy of past cold conditions is widely recorded and often exceptionally well preserved. Pleistocene glacial features such as cirques, U-shaped valleys, limestone pavements, and moraines often dominate upland landscapes. These are closely associated with karstic processes in limestone uplands. A key advance of the last decade has been the wider application of radiometric dating and this has shown that glaciers were active in many areas during the Middle and Late Pleistocene. Periglacial features such as rock glaciers, scree formations, and frost-shattered bedrock are also key landscape elements in many upland areas.

Conclusions Active glacial and periglacial environments are present in several of the highest mountains in the Mediterranean and cryospheric processes been shown to have major impacts on geomorphological systems and biotic communities throughout the upland zone (above 500 m). However, it is clear that glacial and periglacial environments are becoming increasingly marginal in this region and most glaciers in the Mediterranean mountains appear to be in retreat. There is no documented evidence of sustained glacier advance in recent decades and, given current climatic trends (Chapter 3),

References Ainsworth, W. F. (1842), Travels and Researches in Asia Minor, Mesopotamia, Chaldea and Armenia. 2 vols. J. W. Parker, London. Akçar, N. and Schlüchter, C. (2005), Paleoglaciations in Anatolia: a schematic review and first results. Eiszeitalter und Gegenwart 55: 102–21. Yavuz, V., Ivy-Ochs, S., Kubik, P.W., Vardar, M., and Schlüchter, C. (2007), Paleoglacial records from Kavron Valley, NE Turkey: Field and cosmogenic exposure dating evidence. Quaternary International 164–5, 170–83. Akçar, N., Yavuz, V., Ivy-Ochs, S., Kubik, P.W., Vardar, M., and Schlüchter, C. (2008) A case for a down wasting mountain

Glacial and Periglacial Environments glacier during Termination I, Vercyenik Valley, NE Turkey. Journal of Quaternary Science 23: 273–85. Allen, J. R. M., Brandt, U., Brauer, A., Hubberten, H. W., Huntley, B., Keller, J., Kraml, M., Mackensen, A., Mingram, J., Negendank, J. F. W., Nowaczyk, N. R., Oberhansli, H., Watts, W. A., Wulf, S., and Zolitschka, B. (1999), Rapid environmental changes in southern Europe during the last glacial period. Nature 400: 740–3. Almagià, R. (1918), Trace glaciali nelle montagne dell’Albania. Rivista Geografica Italiana 25: 85–95. Arpat, E. and Özgül, N. (1972), Rock glaciers in Geyik Da˘gi area, central Taurus. Bulletin of the Mineral Research and Exploration Institute of Turkey 78: 28–32. Awad, H. (1963), Some aspects of the geomorphology of Morocco related to the Quaternary climate. The Geographical Journal 129: 129–39. Bailey, G. N., Lewin, J., Macklin, M. G., and Woodward, J. C. (1990), The ‘Older Fill’ of the Voidomatis valley, North-west Greece and its relationship to the Palaeolithic archaeology and glacial history of the region. Journal of Archaeological Science 17: 145–50. Ballais, J.-L. (1983), Moraines et glaciers Quaternaires des Aurès Algérie. 108ème Congrès national de Sociétés savantes, Grenoble, 1983, Géographie, 291–303. Barbier, A. and Cailleux, A. (1950), Glaciaire et périglaciaire dans le Djurdjura occidental Algérie. Comptes Rendus des Séances de l’Académie des Sciences, Paris. Juillet-Décembre 1950, 365–6. Barrère, P. (1963), La période glaciaire dans l’Ouest des Pyrénéés centrales franco-espagnoles. Bulletin de la Société Géologique de France 5: 516–26. Bayari, S., Zreda, M., Çiner, A., Nazik, L., Törk, K., Özyurt, N., Klimchouk, A., and Sarikaya, A. M. (2003), The extent of Pleistocene ice cap, glacial deposits and glaciokarst in the Aladaglar massif: central Taurids range, southern Turkey. XVI INQUA Congress, Reno, Nevada, 22–30 July 2003. Abstracts with Programmes, 40–9: 144. Blumenthal, M. M. (1958), Vom Agri Dag (Ararat) zum Kaçkar Dag. Bergfahrten in nordostanatolischen Glaenzlanden. Die Alpen 34: 125–37. Boenzi, F., Palmentola, G., Sanso, P., and Tromba, F. (1992), Le tracce glaciali del massiccio dello Smolikas (Catena del PindoGrecia). Rivista Geografica Italiana 99: 379–93. Bonaparte, Prince Roland (1890), Les variations périodiques des glaciers Français II. Annuaire Club Alpine Français 17: 423–7. Bordonau, J. (1992), Els complexos glàcios-lacustres relacionats amb el darrer cicle glacial als Pirineus. Dept. de Geologia Dinàmica, Geofísica i Paleontologia, Universitat de Barcelona, Ph.D. thesis. Geoforma Ediciones, Logroño. Brosche, K. U. (1978), Formas actuales y limites inferiores periglaciares en la Peninsula Iberica. Estudos Geograficos 151: 131–61. Calvet, M. (2004), The Quaternary Glaciation of the Pyrenees, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations– Extent and Chronology. Part I: Europe. Elsevier: Amsterdam, 119–28. Chueca, J. and Julián, A. (2004), Relationship between solar radiation and the development and morphology of small cirque glaciers Maladeta mountain massif, central Pyrenees, Spain. Geografiska Annaler 86A: 81–9. (2005), Movement of Besiberris rock glacier, Central Pyrenees, Spain: Data from a 10-year geodetic survey. Arctic, Antarctic and Alpine Research 37: 163–70.

379

Saz, M. A., Creus, C., and López, J. I. (2005), Responses to climatic changes since the Little Ice Age on Maladeta Glacier (Central Pyrenees). Geomorphology, 68: 167–82. Çiner, A. (2004), Turkish Glaciers and Glacial Deposits, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations— Extent and Chronology. Part I: Europe. Elsevier, Amsterdam, 419–29. Deynoux, M., and Çörekcioglu, E. (1999), Hummocky moraines in the Namaras and Susam Valleys, Central Taurids, SW Turkey. Quaternary Science Reviews 18: 659–69. CLIMAP Project Members (1976), The surface of ice-age earth. Science 191: 1131–7. Conchon, O. (1978), Quaternary studies in Corsica France. Quaternary Research 9: 41–53. (1986), Quaternary Glaciations in Corsica. Quaternary Science Reviews 5: 429–32. Couvreur, G. (1966), Les formations périglaciares du Haut Atlas central Marocaine. Revue de Géographie du Maroc 10: 47–50. Cviji´c, J. (1900), L’Époque Glaciaire dans la Péninsule des Balkans. Annales de Géographie 9: 359–72. (1913), The ice age in the Prokletije and surrounding mountains (in Serbian). Glas Srpske Kraljevske Akademije Nauka 91: 1–149. (1917), L’Époque Glaciaire dans la Péninsule Balkanique. Annales de Géographie 26: 189–218. D’Alessandro, L., and Pecci, M. (2001), Rischi da valanga sul Gran sasso d’Itlaia. IV Convegno dei giovani ricercatori di Geologia Applicata, CNR-GNGA, 22–24 Ottobre 1998, Chieti. Memorie della Societa Geologica Italiana 56: 315–20. De Sisti, G., D’Orefice, M., Pecci, M., and Ventura, R. (2003), Geomorphology of the summit area of the Gran Sasso d’Italia (Abruzzo, Italy). Geografia Fisica e Dinamica Quaternaria 26: 125–41. Dansgaard, W., Johnsen, S. J., Clausen, H. B., Dahl-Jensen, D., Gundestrup, N. S., Hammer, C. U., Hvidberg, C. S., Steffensen, J. P., Sveinbjörnsdottir, A. E., Jouzel, J., and Bond, G. (1993), Evidence for general instability of past climate from a 250,000year ice-core record. Nature 364: 218–20. Daveau, S. (1971), La Glaciation de la Serra de Estrela. Finisterra, Revista portuguesa de geografia. Centro de Estudos Geográficos, Faculdade de Letras, University of Lisbon, 6: 5–40. (1978), Le périglaciaire d’altitude au Portugal. Colloque sur le Périglaciaire d’altitude du domaine Méditerranéen et Abords. Université Strasbourg, Strasbourg, 63–78. Diener, C. (1886), Libanon. Grundlinien der physischen Geographie und Geologie von Mittel-Syrien. Vienna. Dramis, F. and Kotarba, A. (1994), Geomorphological evidences of high mountain permafrost in Central Appenines. Geografia Fisica e Dinamica Quaternaria 17: 29–36. Dresch, J. and Raynal, R. (1953), Les formes glaciaires et périglaciaires dans le Moyen Atlas. Comptes Rendus Sommaires des Séances de la Société Géologique de France 195–7. Ehlers, J. (1996), Quaternary and Glacial Geology. Wiley, Chichester. Erinç, S. (1952), Glacial evidences of climatic variations in Turkey. Geografisker Annaler 34: 89–98. Fabre, G. and Maire, R. (1983), Néotectonique et morphologénèse insulaire en Grèce: le massif du Mont Ida (Crète). Méditerranée 2: 39–40. Federici, P. R. (1979), Una ipotesi di cronologia glaciale würmiana, tardo e post- würmiana nell’Appenino Centrale. Geografia Fisica Dinamica Quaternaria 2: 196–202.

380

Philip Hughes and Jamie Woodward

Federici, P. R. (1980), On the Riss glaciation of the Apennines. Zeitschrift für Geomorphologie NS 24: 111–16. and Pappalardo, M. (1995), L’evolutione recente dei ghiacciai delle Alpi Maritime. Geografia Fisica e Dinamica Quaternaria 18: 257–69. and Stefanini, M. C. (2001), Evidences and chronology of the Little Ice Age in the Argentera Massif Italian Maritime Alps. Zeitschrift für Gletscherkunde und Glazialgeologie 37: 35–48. Granger, D. E., Pappalardo, M., Ribolini, A., Sapgnolo, M., and Cyr, A. J. (2008), Exposure age dating and Equilibrium Line Altitude reconstruction of an Egesen moraine in the Maritime Alps, Italy. Boreas 37: 245–53. Fernadez Mosquera, D., Marti, K., Vidal Romani, J. R., and Weigel, A. (2000), Late Pleistocene deglaciation chronology in the NW of the Iberian Peninsula using cosmic-ray produced 21 NE in quartz. Nuclear Instruments and Methods in Physics Research B 172: 832–7. Fisinger, W. and Ribolini, A. (2001), Late glacial to Holocene deglaciation of the Colle Del Vei Bouc-Colle Del Sabbione Area (Argentera massif, Maritime Alps, Italy–France). Geografia Fisica e Dinamica Quaternaria 24: 141–56. Florineth, D. and Schlüchter, C. (2000), Alpine evidence for Atmospheric Circulation Patterns in Europe during the Last Glacial Maximum. Quaternary Research 54: 295–308. Ford, D. and Williams, P. (1989), Karst Geomorphology and Hydrology. Unwin Hyman, London. 601. French, H. M. (1996), The Periglacial Environment. 2nd edn. Harlow: Longman. Frochoso, M. and Castañón, J. C. (1995), Comments on ‘Glaciers in Picos de Europa, Cordillera Cantábrica, northwest Spain’. Journal of Glaciology 41: 430–2. Furlan, D. (1977), The climate of Southeast Europe, in C. C. Wallen (ed.), Climates of Central and Southern Europe. Elsevier, Amsterdam, 185–223. Gabrovec, M. (1998), The Triglav glacier between 1986 and 1998. Geografski zbornik 38: 89–110. Gams, I. (1994), Changes of the Triglav glacier in the 1955–94 period in the light of climatic indicators. Geografski Zbornik 34: 81–117. (2001), Notion and forms of contact karst. Acta Carsologica 30/2: 33–46. García, A. (1996), Contribución al conocimiento glaciar de Sierra Nevada a través de los escritos de los viajeros del siglo XVIII y mediados del XIX, in J. Chacón (ed.), Conferencia Internacional Sierra Nevada. Universidad de Granada - Sierra Nevada 96, Granada, 107–16. García-Ruiz, J. M., Valero-Garcés, B. L., Martí-Bono, C., and González-Sampériz, P. (2003), Asynchroneity of maximum glacier advances in the central Spanish Pyrenees. Journal of Quaternary Science 18: 61–72. Gellatly, A. F., Grove, J. M., and Switsur, V. R. (1992), MidHolocene glacier activity in the Pyrenees. The Holocene 2: 266–70. Smiraglia, C., Grove, J. M., and Latham, R. (1994), Recent variations of Ghiacciaio del Calderone, Abruzzi, Italy. Journal of Glaciology 40: 486–90. Gentileschi, M. L. (1967a), Forme crionivali sul gran Sasso d’Italia. Bollettino della Societa Geografia Italiana 8: 34–61. (1967b), Forme crionivali sulla Maiella. Bollettino della Societa Geografia Italiana 8: 325–50. Giraudi, C. (2003), Middle Pleistocene to Holocene Appenine Glaciation Italy. Il Quaternario 16: 37–48.

(2004), The Apennine Glaciations in Italy, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations—Extent and Chronology. Part I: Europe. Elsevier, Amsterdam, 215–23. (2005), Middle to Late Holocene glacial variations, periglacial processes and alluvial sedimentation on the higher Apennine massifs (Italy). Quaternary Research 64: 176–84. and Frezzotti, M. (1997), Late Pleistocene glacial events in the Central Appenines, Italy. Quaternary Research 48: 280–90. Gómez, A., Schulte, L., García, A., and Palacios, D. (1999), Sobre la eistencia de permafrost en Sierra Nevada. Significado geomorfológico y paleoclimático, in L. Pallí and C. Roqué (eds.), Avances en el estudio del Cuaternario español. AEQUA, Universitat de Girona, 181–6. Palacios, D., Ramos, M., Tanarro, L. M., Schulte, L., and Salvador, F. (2001), Location of permafrost in marginal regions: Corral del Veleta, Sierra Nevada, Spain. Permafrost and Periglacial Processes 12: 93–110. Luengo, E., Tanarro, L. M., Schulte, L., and Ramos, M. (2003), Talus instability in a recent deglaciation area and its relationship to buried ice and snow cover evolution (Picacho del Veleta, Sierra Nevada, Spain). Geografiska Annaler 85A: 165–82. Gómez-Órtiz, A. and Salvador Franch, F. (1998), Procesos periglaciares actuales en montaña mediterránea. Ideas clave, trabajos de campo y resultados en Sierra Nevada, in A. GómezÓrtiz, F. Salvador Franch, L. Schulte, and A. García Navarro (eds.), Procesos Biofisicos Actuales en Medios Fríos. Estudios Recientes, Publicaciones Universidad de Barcelona, Barcelona, 217–34. González-Suárez, J. J. and Alonso, V. (1994), Glaciers in Picos de Europa, Cordillera Cantábrica, northwest Spain. Journal of Glaciology 40: 198–9. Granger, D.E., Spagnolo, M., Federici, P., Pappalardo, M., Ribolini, A. and Cyr, A.J. (2006) Last glacial maximum dated by means of lOBe in the Maritime Alps (Italy). Eos Transactions. American Geophysical Union 87, Fall Meeting Supplement, Abstract H53B-0634. Grove, A. T. (2001) The ‘Little Ice Age’ and its geomorphological consequences in Mediterranean Europe. Climatic Change 48: 121–36. Grove, J. M. (2004), Little Ice Ages: Ancient and Modern. Routledge, London, i and ii. and Gellatly, A. F. (1995), Little Ice Age glacier fluctuations in the Pyrénées. Zeitschrift für Gletscherkunde und Glaziolgeologie 31: 199–206. Güner, Y. and Emre, O. (1983), Erciyes Da˘ginda Pleyistocen Buzulla¸smasi ve Volkanizma ile ili¸skisi. Jeomorfoloji Dergisi 11: 23–34. Gurnell, A. M. and Clark, M. J. (1987), Glacio-fluvial sediment transfer: an alpine perspective. Wiley, Chichester. Hamlin, R. H. B., Woodward, J. C., Black, S., and Macklin, M. G. (2000), Sediment fingerprinting as a tool for interpreting longterm river activity: the Voidomatis basin, Northwest Greece, in I. D. L. Foster (ed.), Tracers in Geomorphology. Wiley, Chichester, 473–501. Hérail, G., Hubschman, J. and Jalut, G. (1986), Quaternary glaciation in the French Pyrenees. Quaternary Science Reviews 5: 397–402. Hewitt, S. (2002), Pleistocene river behaviour and sediment transfer in glaciated catchments of the Mediterranean. Ph.D. thesis, University of Leeds.

Glacial and Periglacial Environments Heybrock, W. (1953), Eiszeitliche Gletscherspuren und heutige Schneeverhältnisse im Zentralgebiet des Hohen Atlas. Zeitschrift für Gletscherkunde und Glazialgeologie 2: 317–21. (1954), Firnverhältnisse auf Korsika. Zeitschrift für Gletscherkunde und Glazialgeologie 3: 75–8. Hughes, P. D. (2004), Quaternary glaciation in the Pindus Mountains, northwest Greece. Ph.D. thesis, University of Cambridge. (2007), Recent behaviour of the Debeli Namet glacier, Durmitor, Montenegro. Earth Surface Processes and Landforms 32: 1593–602. and Woodward, J. C. (2008), Timing of glaciation in the Mediterranean mountains during the last cold stage. Journal of Quaternary Science 23: 575–88. (2004), Quaternary glaciation in the Atlas Mountains, North Africa, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciation—Extent and Chronology, iii. Asia, Latin America, Africa, Australia, Antarctica. Elsevier, Amsterdam, 255–60. (2005), Quaternary glacial records in mountain regions: A formal stratigraphical approach.Episodes 28: 85–92. (2006a), Quaternary glacial history of the Mediterranean mountains. Progress in Physical Geography 30: 334–64. Woodward, J. C., Gibbard, P. L., Macklin, M. G., Gilmour, M. A., and Smith, G. R. (2006b), Quaternary glacial stratigraphy and geochronology of the Pindus Mountains, Greece. Journal of Geology 114: 413–34. (2006c), The last glaciers of Greece. Zeitschrift für Geomorphologie 50: 37–61. (2006d), Late Pleistocene glaciers and climate in the Mediterranean region. Global and Planetary Change 50: 83–98. Imbrie, J., Hays, J. D., Martinson, D. G., McIntyre, A., Mix, A. C., Morley, J. J., Pisias, N. G., Prell, W. L., and Shackleton, N. J. (1984), The orbital theory of Pleistocene climate: support from a revised chronology of the marine 18 O record, in A. Berger, J. Imbrie, G. Hays, G. Kukla, and B. Saltzman (eds.), Milankovitch and Climate. Reidel, Dordrecht, 269–306. Jalut, G., Montserrat, J., Fontunge, M., Delibrias, G., Vilaplana, J., and Juliá, R. (1992), Glacial to interglacial vegetation changes in the northern and southern Pyrenees: deglaciation, vegetation cover and chronology. Quaternary Science Reviews 11: 449–80. Jiménez-Sánchez, M. and Farias, P. (2002), New radiometric and geomorphologic evidence of Last Glacial maximum older than 18 ka in SW European Mountains: the example of Redes Natural Park Cantabrian Mountains, NW Spain. Geodinamica Acta 15: 93–101. Jones, A. P. (2000), Late Quaternary sediment sources, storage and transfers within mountain basins using clast lithological analysis: Pineta Basin, central Pyrenees, Spain. Geomorphology 34: 145–61. Karkanas, P. (2001), Site formation processes in Theopetra cave: a record of climatic change during the Late Pleistocene and early Holocene in Thessaly, Greece. Geoarchaeology 16: 373–99. Kelletat, D. (1969), Verbreiung und Vergesellschaftung Rezenter Periglazialerscheinungen im Apennin. Goettinger Geographische Abhandlungen 48: 1–114. Kotarba, A., Hercman, H., and Dramis, F. (2001), On the age of Campo Imperatore glaciations, Gran Sasso Massif,

381

Central Italy. Geografia Fisica e Dinamica Quaternaria 24: 65–9. Krajick, K. (2004), All downhill from here? Science 303: 1600–2. Kuhleman, J., Frisch, W., Székely, B., Dunkl, I., Danišík, M., and Krumei, I. (2005), Würmian maximum glaciation in Corsica. Austrian Journal of Earth Sciences 97: 68–81. Kurter, A. (1991), Glaciers of the Middle East and Africa–Glaciers of Turkey, in R. S. Williams and J. G., Ferrigno (eds.), Satellite Image Atlas of Glaciers of the World. United States Geological Survey Professional Paper 1386-G-1: 130. and Sungur, K. (1980), Present glaciation in Turkey. International Association of Hydrological Sciences 126: 155–60. Lamb, H. H. (1977), Climate: Present, Past and Future, ii. Climatic History and the Future. Methuen, London. Lebourg, T., Fabre, R., Clement, B., and Frappa, M. (2003), Highmountain landslides in the Atlantic Pyrenees: their relationship with the geology and geomorphology. Bulletin of Enginneering Geology and the Environment 62: 221–9. Letsch, K. (1956), Firnverhältnisse auf Korsika. Zeitschrift für Gletscherkunde und Glazialgeologie 3: 268. Lewin, J., Macklin, M. G., and Woodward, J. C. (1991), Late Quaternary fluvial sedimentation in the Voidomatis basin, Epirus, northwest Greece. Quaternary Research 35: 103–15. Liedtke, H. (1962), Eisrand und Karstpoljen am Westrand Lukavica-Hochfläche. Erdkunde 16: 289–98. Louis, H. (1926), Glazialmorphologishche Beobachtungen im albanischen Epirus. Zeitschrift der Gesellschaft für Erdkunde 398–409. Magaš, D. (2002, Natural-geographic characteristics of the Boka Kotorska area as the basis of development. Geoadria 7/1: 51–81. Marino, A. (1992), Nota preliminaire sul fenomeno glaciologico della Gretta del Gelo (Monte Etna). Geografia Fisica Dinamica Quaternaria 15: 127–32. Marjanac, L. and Marjanac, T. (2004), Glacial history of the Croatian Adriatic and Coastal Dinarides, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciation—Extent and Chronology. Part I: Europe. Elsevier, Amsterdam, 19–26. Marovic, M. and Markovic, M. (1972), Glacial morphology of the Durmitor Mt. wider area (in Serbian). Geoloski anali Balkanskog poluostrva 37: 37–48. Martinson, D. G., Pisias, N. G., Hays, J. D., Imbrie, J., Moore, T. C., and Shackleton, N. J. (1987), Age dating and the orbital theory of the ice ages: development of a high resolution 0– 300,000 year chronostratigraphy. Quaternary Research 27: 1–29. Menkovic, L., Markovic, M., Cupkovic, T., Pavlovic, R., Trivic, B., and Banjac, N. (2004), Glacial Morphology of Serbia Yugoslavia, with comments on the Pleistocene Glaciation of Montenegro, Macedonia and Albania, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations— Extent and Chronology. Part I: Europe. Elsevier, Amsterdam, 379–84. Messerli, B. (1967), Die eiszeitliche und die gegenwartige Vertgletscherung im Mittelemeeraum. Geographica Helvetica 22: 105–228. (1980), Mountain glaciers in the Mediterranean area and in Africa. World Glacier Inventory. World Glacier Inventory, International Association of Hydrological Sciences 126: 197–211. Michelier, P. L. (1887), Rapport sur les variations des glaciers des Pyrénées. Annales du bureau Central Météorologique du France 1: 1–235.

382

Philip Hughes and Jamie Woodward

Mistardis, G. (1952), Recherches glaciologiques dans les parties supérieures des Monts Oeta et Oxya Grèce Centrale. Zeitschrift für Gletscherkunde und Glazialgeologie 2: 72–9. Nicod, J. (1968), Premières recherches de morphologie karstique dans le massif du Durmitor. Meditéranée 3: 187–216. Niculescu, C. (1915), Sur les traces de glaciation dans le massif Smolica chaîne du Pinde méridional. Bulletin de la Section Scientifique de l’Academie Roumaine 3: 146–51. Ohmura, A., Kasser, P., and Funk, M. (1992), Climate at the equilibrium line of glaciers. Journal of Glaciology 38: 397–411. Palacios, D., de Andrés, N., and Luengo, E. (2003), Distribution and effectiveness of nivation in Mediterranean mountains: Peñalara (Spain). Geomorphology 54: 157–78. Marcos, J., Andres, N., and Vazquez, L. (2007), Last glacial maximum and deglaciation in central Spanish mountains. Geophysical Research Abstracts 9: 05634. Palgrave, W. G. (1872), Vestiges of the glacial period in northeastern Anatolia. Nature 5: 444–5. Pallàs, R., Rodés, Á., Braucher, R., Carcaillet, J., Ortuño, M., Bordonau, J., Bourlès, D., Vilaplana, J. M., Masana, E., and Santanach, P. (2007), Late Pleistocene and Holocene deglaciation in the Pyrenees: a critical review and new evidence from 10 Be ages, south-central Pyrenees. Quaternary Science Reviews 25: 2937–63. Palmentola, G., Boenzi, F., Mastronuzzi, G., and Tromba, F. (1990), Osservazioni sulle tracce glaciali del M. Timfi, catena del Pindo (Grecia). Geografia Fisica e Dinamica Quaternaria 13: 165–70. Baboci, K., Gruda, G., and Zito, G. (1995), A note on rock glaciers in the Albanian Alps. Permafrost and Periglacial Processes 6: 251–7. Pappalardo, M. (1999), Remarks on the present-day condition of the glaciers in the Italian Appenines. Geografia Fisica e Dinamica Quaternaria 18: 257–69. Pechoux, P. Y. (1970), Traces of glacial action in the Mountains of Central Greece. Revue de Géographie Alpine 58: 211–24. Penck, A. (1885), La Période glaciaire dans les Pyrénées. Bulletin de la Societe d’Histoire Naturelle de Toulouse 19: 105–200. (1900), Die Eiszeit auf der Balkanhalbinsel. Globus 78: 133–78. Ponel, P., Andrieu-Ponel, V., Parchoux, F., Juhasz, I., and De Beaulieu, J.-L. (2001), Late-glacial and Holocene highaltitude environmental changes in Vallée des Merveilles AlpesMaritimes, France: insect evidence. Journal of Quaternary Science 16: 795–812. Poser, J. (1957), Klimamorphologische Probleme auf Kreta. Zeitschrift für Geomorphologie 2: 113–42. Pumpelly, R. (1859), Sur quelques glaciers dans l’île de Corse. Bulletin de la Société Géologique de France 17: 78. Raynal, R. (1952), Quelques examples de l’action du froid et le neige sur les formes du relief au Maroc. Notes Marocaines 2: 14–18. Dresch, J., and Joly, F. (1956), Deux exemples régionaux de glaciation Quaternaire au Maroc: Haut Atlas Oriental, Moyen Atlas Septentrional. IV Congrès INQUA, Rome-Pisa, 1953 105–17. Ribolini, A. (1996), Note geomorfologiche sull’alta Valle del Sabbione e sulla Val’d’Ischietto (Gruppo dell’Argentera, Alpi Marittime). Geografia Fisica e Dinamica Quaternaria 19: 79–91.

(1999), Areal distribution of rock glaciers in the Argentera massif (Maritime Alps) as a tool for recent glacial evolution reconstruction. Geografia Fisica e Dinamica Quaternaria 22: 83–6. and Fabre, D. (2006), Permafrost existence in rock glaciers of the Argentera Massif, Maritime Alps, Italy. Permafrost and Periglacial Processes 17: 49–63. Robinson, D. A. and Williams, R. B. G. (1992), Sandstone weathering in the High Atlas, Morocco. Zeischrift für Geomorphologie 36: 413–29. Roth von Telegd, K. (1923), Das albanisch-montenegrinische Grenzgebiet bei Plav (Mit besonderer Berücksichtigung der Glazialspuren), in E. Nowack (ed.), Beiträge zur Geologie von Albanien. Neues Jahrbuch für Mineralogie 1. Schweizerbart, Stuttgart, 422–94. Sahsamanoglu, H. S. (1989), Mount Olympus Summer Snowfall. International Journal of Climatology 9: 309–19. Sancho, L. G., Palacios, D., De-Marcos, J., and Valladares, F. (2001), Geomorphological significance of lichen colonization in a present snow hollow: Hoya del cuchillar de las navajas, Sierra de Gredos (Spain). Catena 43: 323–40. Sarıkaya, A. M., Çiner, A., and Zreda, M. (2003), Late Quaternary glaciation of Erciyes volcano, central Turkey. XVI INQUA Congress, Reno, Nevada 23–30 July 2003, Abstracts with Programmes, Abstract 40–4: 144. Zreda, M., Çiner, A., and Zweck, C. (2008), Cold and wet Last Glacial Maximum on Mount Sandlras, SW Turkey, inferred from cosmogenic dating and glacier modelling. Quaternary Science Reviews 27: 769–80. Sawicki, R. von. (1911), Die eiszeitliche Vergletscherung des Orjen in Süddalmatien. Zeitschrift für Gletscherkunde 5: 339–50. Seret, G., Dricot, J., and Wansard, G. (1990), Evidence for an early glacial maximum in the French Vosges during the last glacial cycle. Nature 346: 453–6. Serrano, E., Agudo, C., and Pison, E. M. (1999), Rock glaciers in the Pyrenees. Permafrost and Periglacial Processes 10: 101–6. and González Trueba, J. J. (2002), La deglaciación de la laalta montaña. Morphología, evolución y fases morfogenéticas glaciares en el macizo del Posets (Pirineo Aragonés). Revista Cuaternario y Geomorfologia 16: 111–26. Serrat, D. (1979) Rock glaciers and moraine deposits in the eastern Pyrenees, in C. Schlüchter (ed.), Moraines and Varves. Balkema, Rotterdam, 93–100. and Ventura, J. (1993), Glaciers of the Pyrenees, Spain and France, in R. S. Williams and J. G. Ferrigno (eds.), Satellite Image Atlas of Glaciers of the World. United States Geological Survey Professional Paper 1386-E-2: 49–61. Sestini, A. (1933), Tracce glaciali sul Pindo epirota. Bollettino della Reale Società Geografica Italiano 10: 136–56. Sibrava, V., Bowen, D. Q., and Richmond, G. M. (eds.) (1986), Quaternary Glaciations in the Northern Hemisphere. Quaternary Science Reviews 5: 1–511. Sifrer, M. (1963) New findings about the glaciation of Triglav. Geografiski zbornik 8: 157–210. Simon, M., Garcia, I., Cabezas, O., Sanchez, S., and Gómez-Ortiz, A. (1994), Terrenos configurados ordenados en la alta montana mediterranea. Pirineos 144: 71–85. Smart, P. L. (1986), Origin and development of glacio-karst closed depressions in the Picos de Europa, Spain. Zeitschrift für Geomorphologie NS 30: 423–43.

Glacial and Periglacial Environments Smith, G. W., Nance, R. D., and Genes, A. N. (1997), Quaternary Glacial History of Mount Olympus. Geological Society of America Bulletin 109: 809–24. Smith, K. (2004), Trekking in the Atlas Mountains. Cicerone, Milnthorpe. Toro, M., Flower, R. J., Rose, N. L., and Stevenson, A. C. (1993), The sedimentary record of the recent history in a high mountain lake in central Spain. Verhandlungen Internationale Vereinigung Limnologie 25: 1108–12. Tzedakis, P. C. (1994), Vegetation change through glacialinterglacial cycles: a long pollen sequence perspective. Philosophical Transactions of the Royal Society of London B345: 403–32. Lawson, I. T., Frogley M. R., Hewitt G. M., and Preece R. C. (2002) Buffered Tree Population Changes in a Quaternary Refugium: Evolutionary Implications. Science 297: 2044–7. Frogley, M. R., Lawson, I. T., Preece, R. C., Cacho, I., and de Abreu, L. (2004), Ecological thresholds and patterns of millennial-scale climate variability: The response of vegetation in Greece during the last glacial period. Geology 32: 109–12. Vieira, G., Ferreira, A. B., Mycielska-Dowgiallo, E., Woronko, B., and Olszak, I. (2001), Thermoluminescence Dating of Fluvioglacial Sediments Serra da Estrela, Portugal. V REQUI – I CQPLI, Lisbon, 23–7 July 2001, 85–92. Mora, C., and Ramos, M. (2003), Ground temperature regimes and geomorphological implications in a Mediterranean mountain (Serra da Estrela, Portugal). Geomorphology 52: 57–72. Waltham, A. C. (1978), The Caves and Karst of Astraka, Greece. Transactions of the British Cave Research Association 5: 1–12. Woodward, J. C. (1997), Late Pleistocene rockshelter sedimentation at Megalakkos, in G. N. Bailey (ed.), Klithi: Palaeolithic

383

Settlement and Quaternary Landscapes in Northwest Greece, ii. Klithi in its Local and Regional Setting. MacDonald Institute for Archaeological Research, Cambridge, 377–93. Lewin, J., and Macklin, M. G. (1992), Alluvial sediment sources in a glaciated catchment: the Voidomatis basin, northwest Greece. Earth Surface Processes and Landforms 16: 205–16. (1995), Glaciation, river behaviour and the Palaeolithic settlement of upland northwest Greece, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 115–29. and Smith, G. R. (2004), Pleistocene Glaciation in the Mountains of Greece, in J. Ehlers and P. L. Gibbard (eds.), Quaternary Glaciations—Extent and Chronology. Part I: Europe. Elsevier, Amsterdam, 155–73. Hamlin, R. H. B., Macklin, M. G., Hughes, P. D., and Lewin, J. (2008), Glacial activity and catchment dynamics in northwest Greece: Long-term river behaviour and the slackwater sediment record for the last glacial to interglacial transition. Geomorphology 101: 44–67. World Glacier Monitoring Service (2003), Glacier Mass Balance Bulletin 7 , accessed 23 October 2008. (2005), Glacier Mass Balance Bulletin 8 , accessed 23 October 2008. Wright, H. E. (1962), Pleistocene glaciation in Kurdistan. Eiszeitalter und Gegenwart 12: 131–64. Xoplaki, E., Maheras, P., and Luterbacher, J. (2001), Variability of climate in meridional Balkans during the periods 1675–1715 and 1780–1830 and its impact on human life. Climatic Change 48: 581–615. Yılmaz, Y., Güner, Y., and S¸ aro˘glu, F. (1998), Geology of the Quaternary volcanic centres of the east Anatolia. Journal of Volcanology and Geothermal Research 85: 173–210.

This chapter should be cited as follows Hughes, P. D. and Woodward, J. C. (2009), Glacial and periglacial environments, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 353–383.

This page intentionally left blank

13

Coastal Geomorphology and Sea-Level Change Iain Stewart and Christophe Morhange

Introduction The intricate shores of the Mediterranean Sea twist and turn for some 46,000 km, with three-quarters of their convoluted length confined to only four countries— Italy, Croatia, Greece, and Turkey. Just over half the coast is rocky, much of it limestone, with the remainder encompassing almost every type of littoral environment (exceptions being coral reefs and mangrove wetlands) (Table 13.1). Such littoral diversity has long made the seaboard of southern Europe, the Levant, and North Africa a fruitful natural laboratory for studying coastal geomorphology and sea-level change. The virtually enclosed sea ensures that wave processes are generally modest and the tidal range is limited (often less than half a metre), a combination that permits observational evidence of many modern shoreline features to be related precisely to mean sea level. Consequently, relative shifts in the position of now relict coastal features can be used to track the rhythms of relative sea-level change and shoreline evolution. Such rhythms have a bearing on several aspects beyond the physical geography of the Mediterranean basin: they inform archaeological reconstructions of the past settlement and exploitation of a coastal zone that has been an important focus of human activity since Palaeolithic times; they provide testing and fine-tuning for geophysical, geodynamic, and palaeoclimatic models for the region; and they set the backdrop to contemporary societal issues, such as future sea-level rise and coastline adjustments to mass tourism, which threaten the long-term sustainability of the Mediterranean littoral. In this chapter, we review these diverse facets of

the Mediterranean coastal realm to provide a synthesis of how these shores have evolved into their present-day appearance.

Morphotectonics of the Mediterranean Seaboard The Mediterranean occupies the convergence zone between two major tectonic plates, Africa and Europe, with a third, Arabia, pressing from the east. Caught within the collisional vice of these great plates are several minor plates and crustal blocks, most notably Anatolia and Apulia. The result is a complex network of plate tectonic structures that define the general configuration of the seaboard (Figure 13.1). In particular, two major subduction systems partition the Mediterranean basin into a patchwork of minor basins and subsidiary seas (Krijgsman 2002; Chapter 1). In the eastern Mediterranean, the Hellenic arc subduction system and its former extension towards Cyprus have advanced southwards consuming the ancient Tethyan ocean floor of the Ionian and Levantine Seas, and in doing so has stretched open a major new seaway, the Aegean, in its wake. In the western Mediterranean, the Calabrian arc subduction system has similarly migrated south-eastwards destroying old Tethyan ocean floor ahead of it and rifting open a suite of successive young marine basins behind (the Alboran, Valencia, Balearic-Algerian, and Tyrrhenian) (Figure 13.1). Only the Adriatic is neither ancient ocean nor young rift, but instead is a drowned epicontinental

386

Iain Stewart and Christophe Morhange

TABLE 13.1. Coastal environments around the Mediterranean Sea classified into bedrock coasts and accretion coasts (which include beaches, dunes, marshes, lagoons, estuaries, and deltas)

shores of restricted seas enclosed by major land masses and island chains.

Country

Littoral Cells

Accretion (km)

%

Spain France Italy Malta Yugoslavia Albania Greece Turkey Cyprus Syria Lebanon Israel Egypt Libya Tunisia Algeria Morocco

Bedrock (km) 80 1,090 3,181 180 4,893 125 10,500 3,115 391 119 146 10 50 90 260 600 256

% 3 64 40 100 80 30 70 60 50 65 65 5 5 5 20 50 50

2,370 613 4,772 0 1,223 293 4,500 2,076 391 64 79 190 900 1,680 1,040 600 256

92 36 60 0 20 70 30 40 50 35 35 95 95 95 80 50 50

Total North coast South and east coasts

25,086 23,164

54 59

21,047 15,847

46 41

1,922

27

5,200

73

Note: More than 5% of Spain’s coastline may be described as artificial. Yugoslavia includes data for Montenegro, Croatia, and Slovenia. The north coast is defined as Spain to Turkey in the data set above (and includes all the major islands for each country) and the south and east coastline is Syria to Morocco (including Cyprus). Total length of coastline = 46,133 km. Source: Modified from Grenon and Batisse (1989).

platform. In geodynamic terms, therefore, the coastlines on and immediately inboard of the Hellenic and Calabrian arcs are the tectonically mobile borders of young active basins, while those around the periphery of the Mediterranean are, by and large, tectonically stable vestiges of the Tethyan passive margin (Figure 13.1; Chapter 1). The geodynamic complexity of the Mediterranean ensures that the three main tectonic types of coastline coexist here: collision, trailing-edge, and marginalsea coasts (Inman and Nordstrom 1971; Inman 1994; Davis 1996). Collision coasts, which characterize the narrow-shelf, mountainous seaboards of plates colliding with each other are arguably best typified by the Strait of Gibraltar collision zone where the African and European plates directly impinge. Trailing-edge coasts, which develop as wide-shelf plains on the rifted flanks of continents, characterize much of the gently warped and foundered North African margin. However, in general, the strong and pervasive tectonic partitioning in the Mediterranean means that its coastal configuration is best viewed as a nested set of marginal-sea coasts— narrow shelves fronting steep hinterlands along the

The fundamental units of study for coastal evolution, littoral cells, correspond to coastal compartments that delineate complete systems of sediment sources, transport paths, and sinks and within whose boundaries the budget of sediment is balanced (Carter 1988). The sediment dynamics of the Mediterranean littoral are strongly related to the size and character of these marginal-sea coasts. Such coasts, because they front onto smaller water bodies, are typically characterized by more limited fetch and reduced swell dynamics. In these settings, river deltas (Figure 13.2) become especially prominent and serve as important sources of sediment for littoral cells. In the southern Mediterranean, marginal-sea coasts typically exhibit large sedimentation cells, some up to hundreds of kilometres in length. Indeed, one of the world’s largest littoral cells stretches 700 km from Alexandria on the Nile delta to offshore northern Israel (Figure 13.3). Within this Nile littoral cell, sediment is swept eastwards from the delta mouth, 1 million m3 yr−1 moving by wave-dominated longshore transport and 10 million m3 yr−1 carried by coastal currents of the east Mediterranean gyre (Inman and Jenkins 1984; Chapter 2). This eastward drift of sediment is locally interrupted by eddy currents at prominent headlands such as the Damietta promontory, temporarily entraining sediment in migrating offshore sandy shoals or ribbons (Murray et al. 1981). Further east, the gradual northerly bend into the Levant coastline produces a progressive divergence in sediment flow, with the shallow longshore component becoming increasingly susceptible to wind action and feeding the extensive sand-dunefields along the coasts of the delta, Sinai, Gaza, and Israel (the ‘dry’ sink), while the shelf currents funnel the bulk into the Akhviz Submarine Canyon north of Haifa and into the Levantine basin (the ‘wet’ sink). It is no surprise that this extensive regionally simple sediment routeway has evolved along the long-lived and stable margin of the ancient Tethyan basin. Elsewhere in the Mediterranean, more active and complex geodynamics foster more complicated sedimentation cells. In the west, the Alboran Sea presents a younger and tectonically active basin in which a more restricted and intricate seaway combines with intense surface water gyres (Chapter 2) and strong winds from the Azores high-pressure cell to create smaller and more convoluted sediment transport cells

Coastal Geomorphology and Sea-Level Change

387

(a)

Europe m

/y

r

35

mm

/yr

1 mm/yr

1m

an

uri

Lig

Ad

ria tic

cia

n Vale

BalearicAlgerian

Tyrrhenian

C Ionian

ran Albo

Aegean

H Levantine

ia

Arab

Africa (b)

Fig. 13.1. Major tectonic structures (a) and associated seismicity (b) of the Mediterranean region, highlighting how active geodynamic zones define the general coastal configuration of the region. The coastlines on and immediately inboard of the Hellenic (H) and Calabrian (C) arcs are the tectonically mobile edges of young active basins, while the coasts around the rest of the Mediterranean are generally much more tectonically stable remnants of the Tethyan passive margin (Chapter 1).

388

Iain Stewart and Christophe Morhange

Rhône

Po Arno Ombrone

Ebro

Adriatic

Tiber

Valencia

Axios

Basento Adra

BalearicAlgerian

Tyrrhenian

Sperkhiós

Alboran

Gediz Küçük Menderes Büyük Menderes

Achelóos

Ionian Aegean

Levantine

Sirte

Figure 13.3

Nile

Circulation pattern in the upper Mediterranean water mass Delta Characteristics

Ebro

Rhône

Po

Nile

Plio-Quaternary thickness (m) Holocene thickness Delta area (km2) Continental shelf area (km2) Shelf width (km) Shelfbreak (m) Onset of delta construction Annual discharge (tons/year) historical time Annual discharge (tons/year) modern time Delta-front advance (km/century)

1,000-1,500 30 350 9,000 65 100 early Holocene –– 3.5 × 106 5

1,000-1,500 35 720 9,000 70 100 early Holocene 60 × 106 3 × 106 4

5,000 30 770 20,000 –– 100 early Holocene –– 20 × 106 0.5-7

3,500 modified 22,000 17,000 70 250 early Holocene 140 × 106 –– 1.5

Fig. 13.2. Coastal morphodynamics of the Mediterranean basins showing the general near-surface water circulation pattern (see Chapter 2) and the locations (large triangles) and attributes (tabulated data) of the four major delta shelves: Ebro, Rhône, Po, and Nile (after Got et al. 1985). Small triangles indicate the location of other deltas discussed in the text. The box shows the area in Figure 13.3.

(Goy et al. 2003). Most of the southern coast of Spain trends east–west, roughly parallel to the prevailing winds, and this situation favours longshore currents and littoral drift. However, where prominent bays and headlands break this trend, the sediment transport routes are intercepted and occasionally locally reversed, thereby delimiting second- and third-order littoral cells. Within the even more tectonically fragmented coastal landscape of the circum-Aegean Sea (Figure 13.1), faultbounded gulfs and uplands produce a complex shelf and nearshore bathymetry that creates variable patterns of sediment routing over distances of a few tens of kilometres to kilometres. These often switch in character across active tectonic structures creating a heterogeneity of sediment sources and sinks and a collage of littoral environments (Leeder et al. 1991; Collier et al. 1995).

Tectonics, Climate, and Sea Level As well as partitioning littoral sedimentation systems, geodynamic processes determine whether the Mediterranean shores are emerging or subsiding (Milliman 1992). The fastest geodynamic movements are occurring at rates of 30–40 mm per year horizontally and about 1 or 2 mm per year vertically, which mean that over thousands to millions of years their cumulative effects make dramatic changes to the coastal configuration. The most pervasive effect of this has been the gradual closure of the seaway by Africa– Europe convergence, a process responsible for triggering a regional dessication event, the Messinian Salinity Crisis (Hsü et al. 1973), which in turn predicated the large-scale reshaping of the Mediterranean seaboard (Chapter 1).

Coastal Geomorphology and Sea-Level Change

389

Fig. 13.3. The Nile littoral cell extends along the south-eastern Mediterranean coast from Alexandria, Egypt, to the Akhziv Submarine Canyon, Israel. Sediment transport paths are shown by dark arrows (modified from Inman and Jenkins 1984).

Several million years of subduction-related tectonic emergence of the Betic-Rif area had gradually shallowed the narrow seaways that connected the Mediterranean to the Atlantic Ocean (Krijgsman et al. 1999; Warny et al. 2003; Duggen et al. 2003, 2004). Around 5.96 million years ago, this tectonic uplift, aided perhaps by global glacio-eustatic processes (Clauzon et al. 1996; Hodell et al. 2001), finally closed the Atlantic gateway. Over the course of the next few hundred thousand years, evaporation of the now largely landlocked basin dropped water levels by at least 800 m, and probably as much as 1,300 m, below their modern equivalent (Ben Gai et al. 2005). Thick sequences of evaporites were deposited in the hypersaline abyssal plains. Around the shores of the Mediterranean salt lake the plummeting base-level triggered a major phase of river downcutting, creating vast planation surfaces and carving extensive

canyon networks. The most formidable river, the Nile, cut a canyon that was three times longer than the Grand Canyon with a similar depth and width (Said 1981), and major rock-cut gorges and cataracts are known to exist beneath the Rhône, the Ebro, and the Po rivers. Pronounced coastal progradation at these canyon mouths would form the earliest submarine cones of the region’s great deltas. Today, infilled palaeo-valleys can be found along much of the Mediterranean coast and its continental shelf, vestiges of this Messinian incision (Chapter 1). This ‘great drying’ ended about 5.3 million years ago, when the marine gateway to the Atlantic was restored, due to the westward propagation of the Alboran rift basin (Duggen et al. 2003, 2004) and/or eastward piracy of the Atlantic waters (Blanc 2002; Loget et al. 2005) breaching the Betic-Rifian land bridge at Gibraltar to create the present-day straits.

390

Iain Stewart and Christophe Morhange

Since the Messinian Salinity Crisis, the Mediterranean waters have been maintained through a fragile balance between tectonics and climate, much of it dependent on east–west differences in salinity that reflect a water budget deficit in which outputs from evaporation exceed inputs from precipitation and freshwater influx (Chapters 2 and 8). The long-term effects of this in terms of sea-level changes during Plio-Pleistocene times are poorly constrained, because the Mediterranean lacks both the sensitive marine-continental switches of the neighbouring Red Sea (e.g. Siddall et al. 2004) and the coral-reef staircases, which in locations such as Barbados, the Huon Peninsula, and Tahiti, have yielded long-term eustatic records (for a review, see Lambeck and Chappell 2001). Nevertheless, some constraints on Pleistocene sea-level oscillations have been provided by stratigraphical sequences in subsiding coastal plains and shallow shelves (e.g. van Andel et al. 1990) or in drowned littoral caves (Fornós et al. 2002; Tuccimei et al. 2003; Antonioli et al. 2004). On emerging coastlines, records of sea-level highstands have been derived from flights of marine terraces (e.g. Keraudren and Sorel 1987; Goy and Zazo 1988; Dumas et al. 1993; Carobene and Dai Pra 2003; Zazo et al. 1999, 2003; Rodríguez-Vidal et al. 2004; Ulu˘g et al. 2005).

Coastal Tectonics As well as revealing long-term (104 to 106 year) sealevel records, the ancient shorelines of the Mediterranean can also be used as a measure of geodynamic activity. A particularly useful marker is the shoreline that formed during the Last Interglacial period, 120,000–130,000 years ago, the climatic optimum of Marine Isotope Stage (MIS) 5.5. This episode has left an especially indelible trace in the Mediterranean, in part because global sea levels stood 3–6 m higher than present, and in part because the unusually warm climatic conditions favoured the development of distinctive faunal assemblages. The fossil Strombus bubonius is an index marker of this highstand (Bordoni and Valensise 1998). Figure 13.4a shows the elevation of this geomorphic marker across the Mediterranean realm. In the Strait of Gibraltar, direct continental collision has locally raised the MIS 5.5 terrace to 20 m above sea level, but this drops abruptly back to being within a few metres of modern sea level along most of Spain’s Atlantic and Mediterranean coasts (Zazo et al. 1999, 2003). Other plate-boundary structures coincide with elevated MIS 5.5 terrace elevations, notably where the horizontally slipping North Anatolian Fault intersects the Mar-

mara Sea coastline (Yaltirak et al. 2002), and more significantly in north-eastern Sicily and south-western Calabria (Miyauchi et al. 1994; Bordoni and Valensise 1998; Ferranti et al. 2006), which appear to be on the rise following the Late Pleistocene detachment of the subducted African slab (Westaway 1993). Elsewhere along the Africa–Europe subduction and collisional front, however, MIS 5.5 shorelines are roughly at their original elevation. Perhaps surprisingly, the highest Last Interglacial terraces in the Mediterranean fall not on a plate-bounding but on a prominent interplate rift zone, the Gulf of Corinth fault system (Keraudren and Sorel 1987; Armijo et al. 1996). No other rifted gulfs of the circum-Aegean coast attain such high Last Interglacial shoreline elevations (Kelletat et al. 1976), and along the opposing Turkish Aegean coast no MIS 5.5 terrace sites have so far been reported (Brückner et al. 2004). Along the Mediterranean coast of Turkey, Pirazzoli (1991) attribute higher shorelines to MIS 5.5 but these have not been dated. The general tendency for Last Interglacial marine terraces to lie within a few metres of present-day sea level indicates general long-term land stability throughout most of the circum-Mediterranean region. The marked variability in terrace elevation highlights the role of local (basin-bounding) rather than regional geodynamics. Both these characteristics are also apparent in the Holocene and modern tide-gauge records from the Mediterranean (Flemming 1978, 1993; Emery et al. 1988; Zerbini et al. 1996). According to these recent sea-level data, most of the Mediterranean coast appears to be submergent, although mostly at a low rate (−1.2 mm/yr, Emery et al. 1988). The zones of little or no sea-level change correspond to the coastal tracts between Gibraltar and Genoa in the west and along southern Turkey in the east. In contrast, the most mobile shores are mostly on or immediately inboard of the Hellenic and Calabrian arc areas of the central northern Mediterranean, where they coincide with zones of high earthquake activity (e.g. Pirazzoli et al. 1996a; Stewart et al. 1997; Chapter 16) or active volcanic centres (Dvorak and Mastrolorenzo 1991; Firth et al. 1996; Stiros 2000; Morhange et al. 2006; Chapter 15). The overall picture, therefore, is of considerable variation in both the amount and sense of vertical coastal movements along the Mediterranean littoral.

Isostatic Responses to Ice-Ocean Loading In addition to tectonic deformation, land movements in the Mediterranean littoral are subject to

Fig. 13.4. (a) Elevation of the Last Interglacial (Marine Isotope Substage 5.5) shoreline based on a compilation of published data. Terrace data from Antonioli et al. (2006), Conchon (1975), Kelletat et al. (1976), Bordoni and Valensise (1998), Zazo et al. (1999, 2003) and Yaltirak et al. (2002). Solid lines trace the main geodynamic arcs (Ca = Calabrian, Cy = Cyprus, G = Gibraltar, H = Hellenic), with solid triangles indicating active subduction fronts and open triangles denoting collisional fronts. (b) Rates of late Holocene crustal movement derived from sea-level curves in Pirazzoli (1991) augmented by more recent sea-level studies. Black downward arrows denote subsiding areas and grey upward arrows indicate emerging areas. (c) Predictions of global isostatic adjustment made for Mediterranean tidegauge stations, updated from predictions of Peltier (2001) by adopting the new analysis of Peltier (2004) and using values listed at <www.pol.ac.uk/psmsl/peltier/index.html>, accessed 27 October 2008.

392

Iain Stewart and Christophe Morhange

glacio-hydro-isostatic crustal effects, whereby the marine basin adjusts to past and ongoing fluctuations in surface loads of ice and water. Various numerical models have emerged which make competing predictions on the shoreline responses to ice-ocean loading (Lambeck 1995; Peltier 1998, 2000; Lambeck and Bard 2000). The intricacies of these ice-history/earthrheology models are outside the scope of this chapter, but there are essentially two key contributions to land movement. The first, a glacio-isostatic contribution, comes from rebound of the former ice-mass centres of Europe and North America, whereby mantle material squeezed beneath the Mediterranean crust by ice-sheet depression now flows back causing the previously upwarped Mediterranean ‘forebulge’ to subside. Figure 13.4c shows the predicted effect of this glacial isostatic adjustment in the Mediterranean. The second, a hydro-isostatic contribution, results from meltwater from ice-sheet decay increasing the water load of the global oceans and seas, thereby downwarping the marine basin floors and upwarping their margins (Lambeck et al. 2004a, 2004b). The application of these models to the Mediterranean littoral is discussed in Pirazzoli (2005), but key points are

-115

-120 -120

noted here. In some models, the combined effect of both glacio-isostatic and hydro-isostatic changes is to accentuate general subsidence across much of the central Mediterranean and induce compensatory upwarping of its margins (Lambeck 1995; Lambeck and Purcell 2005). However, according to Peltier (1987) the Mediterranean region is sufficiently remote from the North European centre of glaciation as to be little influenced by the collapse of its forebulge; instead, relative sea-level movements ought to be more significantly influenced by the local water load emplaced on the basin itself. The models also highlight local idiosyncracies. For instance, along parts of the North African and Levantine coast, Lambeck and Purcell (2005) predict that the two isostatic contributions can potentially counteract each other, whilst in the northern Adriatic the Alpine deglaciation probably amplifies the glacial isostatic signal (Lambeck and Purcell 2005). In both areas, accentuated coastal emergence is postulated, making it possible that mid- to late Holocene land uplift may have temporarily outpaced sea-level rise to create locally a highstand shoreline. Figure 13.5 shows sea-level and shoreline predictions for the Mediterranean region for four time slices at 20 ka, 12 ka, 6 ka, and 2 ka

-45

-150 -150

-70 -70

-150 150

-70 -70 -110

-45

-50

-100

-150

-65 -65

-150 -150

(b) 12 ka

(a) 20 ka

-1 -1

0 -1

-8 -8

-1 -5

-8 -8

0

-1 0

-8 -8

(c) 6 ka

0

0

-1 -1 1

(d) 2 ka

Fig. 13.5. Predicted relative sea levels and shorelines across the Mediterranean region at four epochs: a = 20 ka, b = 12 ka, c = 6 ka, and d = 2 ka. The palaeoshoreline positions are defined by the seaward side of the grey shading. For 20 and 12 ka BP, the contour intervals are 5 m. For 6 and 2 ka BP the solid contours denote negative values, the finely dashed contours denote positive values, and the dashed contours correspond to zero change (modified from Lambeck and Purcell 2005).

Coastal Geomorphology and Sea-Level Change

respectively. Note the expansion of the coastal plain at the global LGM when sea level was around 120 m lower than today. In general, however, glacio-isostatic model simulations confirm the findings of most sea-level studies, i.e. that postglacial Mediterranean sea levels have never been higher than at the present day. In other words, any signs of Holocene coastal phenomena found elevated above the modern waterline is strong evidence for local tectonic movements. It is a reminder that geological evidence for past sea levels provide the testing ground for these global isostatic models (e.g. Pirazzoli 1998). Thus, in a comparison between several late Holocene sea-level histories and glacio-hydro-isostatic predictions, Pirazzoli (2005) contends that, in non-tectonic areas, hydro-isostatic effects have been, in places, overestimated (e.g. Sardinia) and in other places underestimated (e.g. southern Tunisia). Pirazzoli (2005) disputes predictions of isostatic subsidence along the Hellenic arc and on the Levant coast. Resolving the discrepancies between model and field data will continue to refine our understanding of Mediterranean postglacial shoreline change, and in the following sections, we discuss how such geologically derived sea-level constraints are determined.

Postglacial Sea-Level Changes Tectonic and glacio-hydro-isostatic effects ensure that there is considerable spatial and temporal variability in land movements within the Mediterranean realm, but the general pattern of postglacial eustatic sea-level rise is adequately known (Lambeck and Chappell 2001). This is known thanks mainly to ‘global’ sea-level curves based upon coral-reef reference sites (e.g. Barbados, the Huon Peninsula, and Tahiti) which show that world sea levels rose by 120–30 m, largely between 16,000 and 8,000 years BP, thereafter stabilizing at present levels by 6,000 years BP (Pirazzoli 1991). However, tracking sea-level behaviour following this mid-Holocene stabilization is more difficult, because corals are low-precision reference markers whose vertical range of repartition in modern reefs is of an equivalent magnitude to eustatic variations over the last 6,000 years (Blanchon and Shaw 1995). This mid- to late Holocene period, however, is critical for isolating various non-eustatic dynamic factors such as isostasy, tectonics, and sea-surface topography that may induce significant local sea-level fluctuations (e.g. Mörner 1996). The Mediterranean littoral, however, offers an opportunity to establish such detailed late Holocene sea-level histories because it combines a diversity of potentially high-

393

precision sea-level proxies with a microtidal regime. As is discussed below, the evidence of former shorelines around the Mediterranean region is key to establishing the local relative sea-level trajectories needed to refine tectonic and glacio-isostatic models.

Evidence of Former Sea Levels in the Mediterranean Littoral Zone A wide range of Mediterranean shoreline indicators can be employed for establishing Holocene sea-level curves ranging from precise ‘sea-level index points’ that securely tie the elevation of the land–sea interface at a specific period in time, to more equivocal and/or less well-dated indicators of former submergence or emergence (Haslett 2000). General reviews are provided by van de Plaasche (1986) and Pirazzoli (1991), but in brief they include a suite of sedimentological, geomorphological, geoarchaeological, and biological proxies.

Sedimentological Proxies Sedimentological evidence provides the most widespread but typically least precise record of shifting shorelines. Fossiliferous sands containing gastropods and bivalves equivalent to those of the modern infralittoral zone can be used, though altitudinal errors may be as large as 10 m. Such metre-scale precision can also be expected from cores from coastal sediment sequences using grain size or fossil fauna to identify contrasting littoral environments (such as terrestrial floodplains, freshwater lakes, seasonal pools, brackish lagoons, beaches, marine delta fronts, and prodelta slopes) and assess the extent of altitudinal change. Even where sharp interfaces between intertidal muds and subaerial peats provide reliable sealevel index points, problems of sediment compaction generally maintain altitudinal uncertainty. In Provence, for example, Vella and Provansal (2000) analysed basal peat formations from the eastern limit of the Rhône delta above a non-deformable substrate and determined an altimetric precision of ±50 cm at best. Clastic shoreline deposits (beachrock) have been used as sea-level indicators where they contain early diagenetic carbonate cements whose textures can be diagnostic of intertidal cementation; though an error of +1 and −5 m may be expected (Fouache et al. 2005b). A potentially better positional accuracy can be obtained from the drowning of littoral caves, whereby speleothems precipitated in emergent caves during lowstands subsequently become encrusted by marine biogenic overgrowths when flooded with sea water (Vesica et al. 2000;

394

Iain Stewart and Christophe Morhange

Fig. 13.6. Schematic representation of a littoral karst cave based on observations on Mallorca Island and Capo Caccia (Sardinia). Note the presence of phreatic overgrowths on speleothems related to former and present-day sea levels (modified from Tuccimei et al. 2003).

Bard et al. 2002; Antonioli et al. 2004; Chapter 10) (Figure 13.6). Complexities include the geometry of the cave system and the fact that there is often a hiatus of several millennia between the end of speleothem formation and marine overprinting caused by the influx of fresh groundwater ahead of saline inundation (Suri´c et al. 2005).

Geomorphological Proxies The geomorphological form of the coast reflects the local balance of bioerosion and bioconstruction. On rocky shores, bioerosion at mean sea-level forms midlittoral notches and tidal platforms. These are best developed on limestone, where they are carved by boring and grazing organisms and dissolved and abraded by wave action. Being a combination of physical, chemical and biological erosion, the nature of the local context strongly controls their precise form (Pirazzoli 1986a; Kershaw and Guo 2001; Naylor and Viles 2002; Trenhaile 2002). The effects of variations of coastal wave energy, for example, produce a continuum of notch forms whose precise relationship to mean sea-level can, in some instances, be difficult to assess (e.g. Kershaw and Guo 2001) (Figure 13.7). Nevertheless, notches and associated platforms are well developed around many carbonate coasts in the Mediterranean, and whether

emergent (e.g. Pirazzoli et al. 1994a, 1996a; Stewart et al. 1997; Rust and Kershaw 2000) or submergent (Faivre and Fouache 2003; Antonioli et al. 2006), they are valuable markers of land movements relative to sea level.

Geoarchaeological Proxies The long human legacy of the Mediterranean means that cultural features are some of the most reliable shoreline markers for reconstructing sea-level changes over recent millenia. Indeed, arguably the best example of the postglacial sea-level highstand comes from Cosquer Cave, southern France, where Palaeolithic wall paintings depicting horses have partially been eroded by the rising sea water, clearly showing that Holocene sea level has never reached higher than its present-day level (Morhange et al. 2001; Figure 13.8). Elsewhere it is archaeological remains that have long provided the reference datum (Flemming 1969, 1978 1993; Blackman 2005): for example, the study of ancient Mediterranean harbours has emerged as an important component of geoarchaeology (Marriner and Morhange 2007). Ancient rock-cut installations such as quarry floors and piscinae (fish ponds) show that the Israeli coast has been tectonically stable for the last two thousand years

Coastal Geomorphology and Sea-Level Change

(a)

395

Wave amplitude

1

2

3

4

Spray Organic rim

High tide Mean sea level Low tide

Very sheltered sites

Very exposed sites

(b)

Fig. 13.7. (a) Effects of variations of coastal-wave energy on marine-notch formation (modified from Rust and Kershaw 2000 and Pirazolli 1986a). 1 and 2 are wave notches formed in quiet conditions at mean sea level; 3 and 4 are surf notches formed in more turbulent conditions as much as 2 m above sea level. (b) A marine notch at Capo Milazzo in north-eastern Sicily illustrating the contrasting morphology of an open (right) to a sheltered (left) position (photo: Iain Stewart).

(Galili and Sharvit 1998; Sivan et al. 2001). Similarly, Lambeck et al. (2004b) have utilized Roman piscinae constructed on or in bedrock along the Tyrrhenian coast of Italy to determine that sea level here two millennia ago was 1–2 m lower than today, a change that they attribute almost entirely to glacio-isostatic subsidence. This result is out of step with previous studies (e.g. Pirazzoli 1976).

Biological Proxies

Present mean sea level

Sublittoral zone

Fig. 13.8. The absence of a Holocene sea level above present datum is supported by the evidence of painted horses on a wall of a half-submerged Palaeolithic cave near Marseilles. Below the current water level of Cosquer Cave the wall paintings are significantly degraded (after Morhange et al. 2001).

The above proxy indicators are by no means independent—ancient harbours, for example, are important stratigraphical archives (e.g. Kraft et al. 2003; Marriner et al. 2006; Marriner et al. 2006)—and many Mediterranean sea-level studies typically combine sedimentological, geomorphological, and archaeological indicators. However, few of these indicators are valuable

396

Iain Stewart and Christophe Morhange

without associated biological indicators. As well as providing often precise reference markers for palaeo-sealevels, biological proxies also serve as dateable deposits from which to establish sea-level histories. Over the last decade or so, the use of biological sea-level indicators in the study of Mediterranean sea-level changes has gradually evolved from a descriptive to a multidisciplinary approach integrating many of the proxies above (Laborel and Laborel-Deguen 1994, 1996). It is an approach based on the recognition that the vertical distribution of the littoral fauna and flora of rocky shores shows a pattern of superimposed ecological belts, a tendency called biological zonation (Péres 1982). According to biological zonation, marine benthic animals and plants are finely adapted to very precise ecological conditions such as light intensity, turbidity, water salinity and temperature, and surf exposure. Consequently, changes in local ecological conditions are followed by a concomitant quantitative and qualitative modification of the organisms with replacement by more tolerant forms. Laborel (1986) discusses this biological zonation in detail and shows how it can be used to measure past sea levels. In very general terms, several parallel zones can be recognized (Figure 13.9) and these are outlined as follows: 1. A supralittoral fringe (or supralittoral zone) never or rarely submerged but wetted by surf in which the biomass is very low and mainly represented by boring endolithic cyanobacteria and grazing gastropods. 2. A midlittoral zone submerged by tides and waves on a regular basis, which displays a pattern of

(a)

parallel algal and faunal belts, with biomass and species diversity increasing downwards. Cyanobacteria, limpets (Patella spp.) and Chitons are the main bio-eroders in this zone. Constructional elements such as the rim-building coralline rhodophyte Lithophyllum byssoides may develop in the north-west Mediterranean. 3. An infralittoral (or sublittoral) zone whose upper limit is marked by a sudden increase in biodiversity (Boudouresque 1971), thus defining a biological sea level that ranges down to the lower limit of marine phanerogams and photophilous algae, i.e. to a mean depth of about 25–35 m. The upper part of this infralittoral zone (also called ‘infralittoral fringe’) is densely populated by brown algae (Cystoseira), Coralline Rhodophytes, fixed vermetid gastropod molluscs (such as Dendropoma sp.), and cirrhipeds, for example Balanus spp. Active erosive agents, such as clionid boring sponges, sea-urchins, and rock-boring mussels (Lithophaga, Hyatella, Coralliophaga spp.), are responsible for rapid underwater erosion of the limestone outcrop. The limit between the midlittoral and the infralittoral corresponds to the ‘biological sea level’(Laborel 1986). The influence of local variations in coastal morphology upon surf exposure explains why this limit may be slightly undulating locally. Aperiodic sea-level oscillations linked to atmospheric pressure or wind variations have little influence upon the marine zonation of living organisms with a lifespan of more than one year. Biological sea level itself is best characterized by

(b) Action

Agent

Erosion construction

Rain water Sea spray Chfthamalus Cyanobacteria, limpets Lythophyllum byssoides Protectio rom Protection from brown algae Sea urchins Clionas Erosion Lithophaga

Resulting morphology

Biological zones

Dissolution karst Supralittoral Biokarst Biokarst Notch Algal rim

Biokarst

Biological zones

Resulting morphology

Supralittoral

Dissolution karst

Upper midlittoral Mid midlittoral

Upper midlittoral Mid midlittoral

Notch

Lower midlittoral

Lower midlittoral

Platform

Subtidal

Biokarst

Subtidal

Erosion Agent construction

Action

Rain water Sea spray

Biokarst

Chfthamalus Cyanobacteria, limpets Dendropoma petraeum Vermetids Sea urchins Clionas

Boring Molluscs Fig. 13.9. A schematic coastal profile showing the main characteristics of bioconstruction and biodestruction on calcareous coasts in (a) the western Mediterranean, and (b) the eastern Mediterranean (modified from Laborel and Laborel-Deguen 1994).

Coastal Geomorphology and Sea-Level Change

the development of the few marine species, with a very narrow depth range, located immediately above (e.g. Lithophyllum byssoides rim) or below (e.g. Dendropoma rim) which are considered the more precise sea-level indicators. When such species are absent, the highest limit of biological perforations by Cliona and Lithophaga (Laborel and Laborel-Deguen 1994) are also excellent proxies. Consequently, littoral algal and vermetid bioconstructions and the upper limits of bioerosive elements (marine burrows and perforations), and fixed invertebrates (oysters, barnacles, solitary vermetids) are commonly used as biological sea-level indicators.

Biological Evidence for the Amplitude and Rapidity of Shoreline Change If a suitable biological indicator is available, and if an accurate study of local ecological conditions has been undertaken, fossil biological sea-level markers can provide reliable constraints on the direction, amplitude, rapidity, complexity, and timing of apparent coastal deformation (ibid.). Of these, indicators of the amplitude and rapidity of relative sea-level displacement are particularly important for discerning tectonic versus eustatic sea-level changes. The amplitude of shoreline change is derived from the altitude of elevated or submerged shorelines. This parameter is best estimated by direct measurement of the altitudinal difference between the upper limit of the elevated (or submerged) remains and the corresponding upper limit of its modern equivalent. This can easily be done with species such as Dendropoma sp. or Lithophyllum byssoides whose populations have a very narrow vertical range closely connected to biological sea level. For species with a wide vertical range such as Lithophaga, balanids, or solitary vermetids, good results are obtained only when the uppermost limits of both fossil and living populations are well delineated or when the fossil remains are correlated with a morphological sea-level indicator such as a tidal notch. Any direct reference to the present instantaneous water level, whether observed or calculated from tide tables, is thus unnecessary and incorrect. For example, the only Mediterranean reefcoral, Cladocora caespitose, can grow from the surface down to more than 40 m, but it does not present any well-marked upper limit nor does it develop distinct sealevel linked build-ups. It should therefore only be used as a proof of submersion but not as a depth proxy and should not be attributed with a precise depth coefficient (Antonioli et al. 1999). When a series of measurements is to be made in a limited area, one must keep in mind as stated above,

397

that a biological sea-level mark is not a perfect horizontal line but may be naturally warped, even over very short distances, by local hydrodynamic variations. For that reason, each individual altitudinal measurement must be done on a vertical profile of its own, including both the fossil specimen and its present equivalent (Laborel 1979). The accuracy that can be obtained depends upon both the state of preservation of the upper limit of fossil populations and local ecological conditions. An accuracy of about ±5 cm was obtained in south-west Crete for a series of remarkably well-preserved vermetid rims (Thommeret et al. 1981). In Euboea, for elevated populations of Lithophaga burrows (Stiros et al. 1992), the lowest accuracies observed (around ±50 cm to several metres) are typically found in relation to Chthamalus populations. Biological sea-level indicators are especially useful in determining the rapidity of relative sea-level changes. In the case of slow uplift (less than a few millimetres per year), biological indicators living in the subtidal zone are killed by emersion and their remains are slowly carried up through the midlittoral zone to the supralittoral zone. Small or fragile skeletons quickly disappear by bioerosion and physical abrasion. Massive algal rims and vermetid constructions are not completely destroyed and can be used as proxies. In contrast, good preservation of fragile elevated remains is the best evidence of a rapid uplift. Determining the rapidity of subsidence is much more difficult than for uplift, because many biological sea-level indicators are either rapidly destroyed by subtidal erosion or covered by new generations of marine organisms. Thus, only bioconstructing species with a very narrow vertical range, such as Lithophyllum byssoides or Dendropoma petraeum., may be of some use. For such bioconstructed rims, it is often possible to obtain valuable information from the distribution of their drowned remnants, provided they are not very old and have not been removed by biological erosion. Eroded remains below the present rim are often found where subsidence is slow. More rapid change would not provide sufficient time for the development of a bioconstructed rim at intermediate depth, and drowned bioconstructions would thus not appear clearly separated (Laborel and LaborelDeguen 1994).

Holocene Sea-Level Histories for Contrasting Mediterranean Coasts As discussed earlier, there is considerable variation in the stability of the Mediterranean coastline, with the

398

Iain Stewart and Christophe Morhange 5000

4000

3000

2000

BC/AD

1000

1000

2000

0 at present 0 NGF 0.1 Lithophyllum lichenoides La Ciotat Laborel et al. 1994

Data from Marseilles

Modern 0.2 irrigation 0.3 channel 0.4

Late Antiquit Antiquity yq quay y

0.5

Archaic quay Archaicquay 0.5 Posts Posts

Roman quay Hellenistic quay Hellenisticquay

1 Late Bronze Age Beach

EarlyBronze Early Bronze Age beach Neolithicbeach Neolithic beach

1.5

2m

Fig. 13.10. Age-depth diagram from Marseilles’s archaeological excavations compared with dated algal rims from nearby rocky cliffs (modified from Morhange et al. 2001 and Laborel and Laborel-Deguen 1994). NGF = Nivellement Général de la France.

geodynamically active Calabrian and Aegean seaboards being especially mobile, the rifted margins of the southern, western, and eastern sectors being largely stable, and localized volcanic centres experiencing irregular paroxysms (Figure 13.4b). In the following section, we examine in more detail the relative sea-level histories established for three sites in these contrasting settings.

‘Stable’ Coast The site of Marseilles (southern France) is a good example of relative sea-level ‘stabilization’ in a socalled tectonically ‘stable’ area (Morhange et al. 2001). Allied with recent archaeological excavations of the ancient harbour, biological indicators have yielded highprecision data for the past 5,000 years (Figure 13.10). One of the best biological sea-level indicators used here is the upper limit of barnacles (Balanus sp.). They commonly develop upon quay walls in clear or polluted waters, stopping abruptly at biological sea level. When their upper limit is continuous, a precision of plus or minus a few centimetres can be obtained. Wherever barnacle-bearing hard surfaces were not available, the upper-limit of subtidal beach onlap layers was used as a sea-level indicator with a precision of ±0.2 m.

Data obtained in Marseilles fit well with indicators from rocky coasts such as Lithophyllum rims (Laborel and Laborel-Deguen 1994). The age-depth diagram shows a regular sea-level rise up to about AD 500 followed by a period of stability. Total rise has been less than 1.5 m since 4,500 years cal. BP. Observations do not show any Holocene level higher than present (Figure 13.10). The rate of mean sea-level rise was 0.4 mm/yr between 4,500 years cal. BP and AD 500, and 0.2 mm/yr thereafter. During the twentieth century, the rate of sea-level rise increased to c.1.5 mm/yr, most likely in connection with global warming. In other words, in settings where geodynamic activity is low or insignificant, background eustatic trends can be isolated.

Volcanic Coasts More complicated coastal elevation changes can occur in areas affected by volcano-tectonic deformation (e.g. Firth et al. 1996). The most detailed record of the complex mobility of volcanically active shores can be shown by the archaeological ruins of Pozzuoli on the Phlegrean Fields in the Bay of Naples. The bioeroded columns of the port’s Roman market were made famous as the frontispiece of Charles Lyell’s Principles of Geology, and

Coastal Geomorphology and Sea-Level Change

3000

Measured curve (Osservatorio Vesuviano)

2000

1/82

1/89

2000

Submersion between 1985 and 1995

Estimated curve (biological indicators) Biological zero 1995

1000

1000

0

0 -1000 1970

1975 1980

1985

1990

1995

Fig. 13.11. Measured relative sea-level changes in the old harbour of Pozzuoli compared to estimated relative sea-level changes using biological indicators (modified from Morhange et al. 2006).

since that time have served as a palaeo-tide gauge to track the deformation history for the area (Dvorak and Mastrolorenzo 1991; Orsi et al. 1999; Morhange et al. 2006). Relative sea-level movements at Pozzuoli are far more intense than the average 50-cm sea-level rise recorded in the north-west Mediterranean sea since Roman times (Pirazzoli 1976). Moreover, the post-Roman sea-level history reveals three cycles of submersion and emergence (Figure 13.11). The first submersion accompanied a period of marine transgression which ended around AD 400–550 without any volcanic eruption. A second submersion affected the site in the early Middle Ages, around AD 700–900, though again no eruptive activity followed this inundation. The third submersion occurred in the late Middle Ages, when it was followed by a well-documented pulse of land uplift that culminated in a major eruption in 1538. More recently, the two major events of 1969–72 and 1982–4 resulted in a total uplift of c.3.5 m, although no eruption occurred during these periods. Indeed, during the last 2,000 years, noneruptive coastal uplift episodes have been the rule rather than the exception. Clearly, here, the dominant sea-level change signal is linked to phases of volcanic activity and quiescence (see Chapter 15).

Pirazzoli et al. 1994a, 1994b, 1996a; Stewart 1996; Stewart et al. 1997). Such ‘coastal palaeoseismology’ studies reveal that seismically active coasts in the Mediterranean realm typically experience decimetreto metre-scale seismic jerks every few centuries to millennia. Arguably the most remarkable example of this seismotectonic action has been reported from the shores of western Crete and nearby Antikythira island, where a series of elevated stepped shorelines are carved into the limestone cliffs supporting well-preserved Vermetid formations. The highest of these contains very fragile shell skeletons that would have been rapidly destroyed in the surf zone (Thommeret et al. 1981). The sea-level history derived from these shorelines is especially complex (Figure 13.12). It reveals that, between 4,000 and 1,700 years BP, ten successive increments of small but abrupt subsidence affected a huge block of lithosphere about 150 km long without noticeable tilting. This phase of subsidence was followed about 1,530 years BP by an abrupt uplift, by about 9 m, of the south-west corner of Crete and a north-eastward tilting of the western

GREECE

South Aegean Sea Antikythira Crete

100 km +3

II III IIIA

I

IV

Height (metres)

3000 1/85

Measured altitudes (mm) against the1995–6 biological zero

Measured altitudes (mm) against the 1968 tide gauge datum

4000

399

IV

+2

IVA V VI VII

+1

Tectonically Active Coast In contrast to volcanically active coastlines where land movements are discernible over months to years, equivalent shoreline movements on earthquake-prone coasts tend to be instantaneous. Abrupt littoral emergence or submergence in these environments has been widely used to identify past large earthquakes in the Holocene coastal record (e.g. Stiros et al. 1992;

0

Present M.S.L. 4000

3500

3000 2500 Years BP

2000

1500

Fig. 13.12. Recent relative sea-level variations in Antikythira island (Greece), from Pirazzoli (1986b) and Pirazzoli et al. (1996b).

400

Iain Stewart and Christophe Morhange

part of the island. In north-west Crete, the Roman harbour of Phalasarna was raised by over 6 m (Pirazzoli et al. 1992; Chapter 16). Relative sea-level change histories at other coastal sites from western Greece to the Levant coasts also appear to exhibit marked vertical tectonic displacements—some of them several metres in amplitude—between 1,750 and 2,000 years BP (Pirazzoli 1986b; Pirazzoli et al. 1996b). These studies attribute the main movement to a great earthquake in AD 365 (Stiros and Papageorgiou 2001), though recalibration of the original radiocarbon dates on elevated littoral fauna led Price et al. (2002) to shift the main emergence event to the 6th century AD, another period of seismic unrest on Crete (Di Vita 1996). The tectonic mechanism for such dramatic coastal displacements also remains uncertain, but they have been ascribed to an exceptional burst of tectonism that occurred on a regional scale in post-Roman times—the

‘Early Byzantine Tectonic Paroxsym’ (Pirazzoli 1986b; Chapter 16). Nevertheless, the literally Byzantine history of Holocene shoreline development in western Crete highlights the potential complexity of land movements at plate boundary zones.

Tsunamis as a Coastal Process The high seismicity of the Mediterranean region means that much of its coastline is subject to recurrent tsunamis (Soloviev 1990; Chapter 17). Unsurprisingly, the main areas of observed tsunami incidence coincide with the principal earthquake belts (Figure 13.13). However, in addition, volcanic-induced tsunamis affect the Tyrrhenian and Aegean seas, and the steep margins of the north-western Mediterranean basin mean that tsunamigenic submarine slumps are potentially commonplace. The result is that while some coastal zones are more susceptible than others, virtually any

Fig. 13.13. Tsunami activity in the Mediterrenean Sea. Historical data compiled from the following sources: Ambraseys (1962) for the Levant; Soloviev (1990) for the western Mediterranean; Tinti and Maramai (1996) for the French Côte d’Azur and Italy; Papadopoulos (1998) for Greece and the Aegean; Altnok and Ersoy (2000) and Yalçıner et al. (2002) for Turkey. Palaeo-tsunami sites are compiled from Wood (1996) for Mallorca, Tunisia, and Egypt; Mastronuzzi and Sansò (2000, 2004) for Apulia, Italy; Dominey-Howes et al. (2000) for the central Aegean; Minoura et al. (2000) for Crete and western Turkey, and Kelletat and Schellman (2002) for Cyprus. Italicized text indicates the principal tsunamigenic zones, including the Bay of Naples (BoN), Aeolian Islands (AI), Messina Straits (MS), Gulf of Corinth (GoC), Santorini (S), North Aegean trough (NAT), and Chios-Izmir zone (C-I). Mega-turbidites are from Rothwell et al. (2000) in the Balearic basin, and Hieke (2000) in the Ionian Sea. Deep-sea homogenite deposits in the eastern Mediterranean are from Cita et al. (1996) and Cita and Aloisi (2000). The lack of recorded tsunamis in some parts of the region, most notably the North African coast, undoubtably reflects a lack of records rather than a real absence of incidence. Data for the Atlantic and the Black Sea are not shown. See Chapter 17.

Coastal Geomorphology and Sea-Level Change

401

TABLE 13.2. Amplitude, duration, permanence, and length of coast affected by various types of rapid relative sea-level change in the Aegean Process or type of change Local tsunami Storm surges Coseismic subsidence or uplift (moderate magnitude earthquakes) Semi-diurnal tides Marine flooding from barrier breaching (natural or artificial coastal barriers)

Amplitude Up to 30 ma 1 mb 0.3–0.8 m (this study) 1–3 m (Gulf of Corinth)c 0.6 m (Gulf of Evvia) Typically 0.1 m or lessd Decimetrese

Duration of change

Permanence of change

Length of coastline affected

Minutes Hours Seconds

<1 hour Hours to days Centuries

c.6 hours

Hours

Tens to hundreds of km Tens to hundreds of km Hundreds of metres to tens of kilometres Entire Aegean region

Hours

Days to centuries

Hundreds of metres

a

Maximum reported value for 1956 south Aegean tsunami (Ambraseys 1960). Value for Z50 extreme sea level based on tide gauge records at N. Halkis, Gulf of Evvia (Tsimplis and Blackman 1997). Based on estimates in Hubert et al. (1996) and Soter (1998). d Zoi-Morou (1981). e Depends on age of barrier, regional sea-level history, coastal morphology, etc. b c

Source: Cundy et al. (2000).

stretch of Mediterranean littoral is prone to tsunami attack. Given their likely historical propensity (Soloviev et al. 2000), tsunamis are expected to be an important agent of abrupt coastal change. Yet detecting and assessing the past impacts of tsunamis on coast evolution is difficult, not least because a number of nearshore mechanisms can cause rapid transient flooding and associated erosion and deposition (Table 13.2) (Cundy 2005). Biostratigraphical studies of coastal sediment sequences have been used to identify likely tsunami flooding events (e.g. Dawson 1996; DomineyHowes et al. 1998), but in many respects their deposits remain indistinguishable from those produced by barrier breaching or storm-surge flooding (Cundy et al. 2000; Dominey-Howes et al. 2000). Furthermore, the long-term (geological) preservation potential of tsunami deposits is poor (Dawson and Stewart 2007). One potentially long-lived coastal expression of large tsunamis, however, may be littoral boulder fields— deposits that require very high-energy emplacement conditions. Mega-clast deposits attributed to tsunami action have been described from both Holocene coastal environments in Apulia (Mastronuzzi and Sansò 2000, 2004), Cyprus (Kelletat and Schellman 2002; Whelan and Kelletat 2002), and Sicily (Monaco et al. 2006), while possible Late Pleistocene examples are described from Egypt, Tunisia, and Mallorca (Wood 2000). The tendency for Mediterranean shores to have experienced sudden and extensive inundation episodes was not limited to past tsunamis. As we discuss in the following section, postglacial transgression resulted in the

abrupt drowning of many low-lying parts of the Mediterranean coastal zone, with major implications for its Palaeolithic and Mesolithic coastal dwellers.

Postglacial Shoreline Development and Human Activity At the height of the global Last Glacial Maximum, a 300-km wide coastal plain was emergent from Tunis to Tripoli, and North Africa was disconnected from the northern Mediterranean only by a few narrow marine straits between Sicily and Tunisia (van Andel 1989). A series of exposed seamounts served as ‘stepping-stones’ within the Strait of Gibraltar (Collina-Girard 2001, 2002), and islands such as Elba, Malta, and most of the Greek Ionian islands were linked to the continent by land bridges. The Gulf of Lions was a continental plain linked with Valencia in present-day Spain, and roughly half the Aegean and Adriatic seas were dry. Between 16,000 and 6,000 years BP, rising postglacial waters inundated continental shelves, submerging low-lying islands, cutting land bridges, and creating many of the islands of the modern Mediterranean. For example, during this period the Greek Cyclades metamorphosed from a wide, low-lying coastal plain into a pair of land masses (one centred on Paros and Andiparos, the other on Mykonos–Tinos–Andros), and then into its present scattered rocky archipelago (Lambeck 1996). The effects were even more striking in the eastern Adriatic, where the flooding of karst topography created one of the

402

Iain Stewart and Christophe Morhange

world’s most indented coasts and a submarine shelf of submerged caves (Chapter 10). The Mediterranean’s coastal caves and plains were important loci for Palaeolithic peoples during the last cold stage (e.g. Gamble 1986), so postglacial coastline changes had dramatic effects on human activity along its shores (Petit-Maire 2003). The broad coastal shelves that constituted the prime Palaeolithic hunting grounds were incrementally submerged, with a probable succession of standstills and spasmodic events (CollinaGirard 1997, 1998; Laborel et al. 1999), so that much of the early coastal human occupation sites of the Mediterranean now lie offshore (Flemming 1998). Pulses of meltwater releases from the great continental ice sheets led to irregular postglacial sea-level rise (Blanchon and Shaw 1995), and consequently episodes of rapid marine encroachment onto coastal plains. The rate at which sea levels rose meant that at times the shores of some coastal lowlands were retreating at several hundred metres per year, forcing seashore communities continually to relocate further inland. The progressive reduction of the relatively hospitable coastal plain environment caused a concomitant loss of resources, squeezed coastal inhabitants into narrower and less conducive littoral zones, and rerouted human traffic patterns (van Andel 1989; Lambeck 1996). Human adaptation to these shifting shores is illustrated by archaeological records from the Late Palaeolithic remains at Franchthi Cave on the Argolid Peninsula of southern Greece (Figure 13.14). At the peak of the last glaciation the coast lay at least 6 km west of Franchthi, but by 12,000–10,000 years BP marine molluscs and small fish bones began appearing in the remains (Jacobsen 1976) providing evidence of an encroaching sea. At times, the rising Mediterranean waters may have had especially dramatic effects. Ryan et al. (1997) argued that the final meltwater rise event around 7,200 years BP was responsible for Mediterranean waters overtopping the narrow land bridge of the Bosphorus Straits and cascading into the freshwater Black Sea basin, inundating burgeoning Neolithic settlements along its shores. The hypothesis of the so-called ‘Black Sea flood’ has been developed further (e.g. Ryan and Pitman 2000; Ryan et al. 2003; Siddall et al. 2004), but the timing, route, and abrupt nature of this marine influx is much disputed (Görür et al. 2001; Aksu et al. 2002; Major et al. 2002; Kerey et al. 2004). Some workers see in this inundation, and others that flooded the various marginal-sea basins, gulfs, and lakes at different times during the postglacial transgression, the root of deluge myths that are deep-seated in

eastern Mediterranean culture (Ryan and Pitman 2000; Petit Maire 2004; Wyatt 2004). However, a more gradual but almost equally dramatic reshaping of the Mediterranean’s coastal geography would take place once the bulk of postglacial sea-level rise had abated.

The Formation and Growth of Mediterranean Deltas Many Mediterranean coasts today comprise long, monotonous sandy beaches interrupted by occasional rocky headlands. Inlets and harbours are rare, except on islands. But this has not always been the case. In antiquity, the shores of the Mediterranean were studded with bays and lagoons, and even large rivers such as the Ebro had harbours at their mouths instead of deltas. The gradual disappearance of this complicated indented coastline may have begun with its submergence by rising postglacial waters, but it was largely accomplished once the last meltwater pulse event had occurred about 8,000 years BP. As sea levels slowly stabilized, sediment supply was able to keep pace with or overwhelm eustatic change, allowing rivers with significant bed loads to establish permanent gravel barrier-beaches and spits around river mouths. Behind these barriers, enormous volumes of clay and silt draped and then buried the drowned glacial coastline. The result was that around the Mediterranean Sea, as on coasts worldwide, deltas began rapidly to advance seawards (Stanley and Warne 1994, 1997). As they grew in extent, the Mediterranean deltas began to subside under their own weight (typically at a few millimetres a year), creating additional space to accommodate ever more sediment. Around the largest deltas, such as the Nile and Rhone, isostatic subsidence, combined with compaction and failure by faulting and slumping, caused harbours to become entombed. The great weight of the Nile delta, for example, has caused its shores to be lowered by 4–5 m in the last 2,500 years (Stanley 1989). Together with localized soft-sediment slumping, this has left Greek port cities such as Herakleion and eastern Canopus entirely submerged in water depths of up to 8 m (Stanley et al. 2004). More often, however, it was simply the advance and lateral migration of the delta fringes that overwhelmed successive bays and inlets with alluvial sediment and transformed them into marshes, wetlands, and floodplains (Chapter 9). In the process, many of the ports and harbours of antiquity gradually lost their connection to the open sea, silted up, and became abandoned (Kraft et al. 2003;

Coastal Geomorphology and Sea-Level Change (b)

14000 20000

8800 5700 Franchthi cave

n M zo oll ne us s ca

s ne bo sh Fi

Neolithic

M m ar ol in lu e sc s

2 km

A ph rch as ae es ol o

N

gi

ca

l

(a)

403

N O

6

D A T A

7

8

9 Cyclope Former shoreline (ages in Cal. yr BP)

Mesolithic

10

Rocky shore 11 (c) Patella

Upper Palaeolithic

Age (thousand years BP)

Cerithium

12

13

14

15 0

20 40 100 200 % of total Total abundance bone remains (number of specimens)

Fig. 13.14. Franchthi Cave in the south-east Peloponnese, Greece showing (a) sea-level transgression since the global Last Glacial Maximum (after van Andel et al. 1980) and (b) the sequence of marine resource exploitation recovered from the archaeological excavations (after Payne 1975 and Shackleton and van Andel 1980). Layout of (a) and (b) after Roberts (1988). (c) Photograph of the site taken in 2006 by Mike Morley. The floor of the cave now lies about 10 m above modern sea level.

Brückner et al. 2004, 2005; Fouache et al. 2005a; Marriner and Morhange 2007). The apparent close association in Classical antiquity between the construction of harbour cities and their demise by delta growth has led many historians and geologists to implicate human-induced denudation of catchment slopes as the root cause of accelerated coastal accretion in late Holocene times. Palaeoenvironmental studies of some ancient harbours clearly detect the anthropogenic influence on sedimentation rates following harbour construction (Morhange

et al. 2003; Marriner et al., 2006). Others, such as Grove and Rackman (2001), however, argue against strong anthropogenic forcing, pointing out that in those coastal lowlands where past human activities have been most intense (e.g. the Anatolian rivers) there is no tendency for the deltas to be disproportionately large, and noting that not all Aegean deltas have slowed their advance in post-Classical times (e.g. Kraft et al. 1977). For some workers, the late Holocene progradation of deltas is a climatic artefact, reflecting a change

Iain Stewart and Christophe Morhange

18

Number of floods

16

25

(a)

(b)

Months with floods

404

14 12 10 8 6 4

20 15 10 5

?

2 0

0 19

19

18

18

17

17

51

01

51

01

51

01

–6

–1

–6

–1

–6

–1

0

0

0

0

0

0

0

0

0

0

–6

–1

–6

–1

51

01

51

01

16

16

15

AD

15

AD D X X II A I V X AD I V X AD IV X D A II X AD X AD I II V D A I AD V

IV

BC

I

BC BC

II

I

II

V

Fig. 13.15. Historical records of coastal flooding for (a) the River Tiber and (b) the River Rhône (see Chapter 18).

towards the present Mediterranean climatic conditions and an increasingly open landscape subject to erosion by episodic storms (Dubar and Anthony 1995; Chapter 11). Many deltas, particularly those in the western and central Mediterranean, experienced their greatest advances during the centuries of climatic deterioration of the Little Ice Age (late sixteenth to early eighteenth centuries AD). Thus Grove and Rackman (2001) report that Rome’s Tiber River delta, which by the Roman period had engulfed the harbour port of Ostia, stalled its further advance until the sixteenth century when frequent bouts of severe flooding (Figure 13.15) drove it forward again. In the seventeenth century the closely charted Po delta advanced at an unparallel rate of 4–9 km along a 30–35-km front (Fabbri 1985). The seventeenth and eighteenth centuries witnessed the Rhône delta also undergo a major advance (50 m annually between 1587 and 1711) and metamorphosis, changing from a meandering to a braided channel and avulsing to its present-day position (Arnaud-Fassetta and Provansal 1999; Vella et al. 2005). Historical records of coastal flooding in the lower reaches of the Rhone and Tiber river systems are shown in Figure 13.15. The Arno river near Pisa and the Spanish deltas of the southern Costas and the Ebro basin also appear to have grown most actively during Little Ice Age times (Chapter 11; Grove and Rackman 2001 and references therein). Few deltas in the east changed significantly during this period, while some in the west, such as the Costa Brava deltas draining the eastern Pyrenees, were similarly unresponsive. The evidence that not all Mediterranean deltas show a marked response to the Little Ice Age sim-

ply underscores the complex and variable pattern of Holocene delta progradation (Grove and Rackman 2001). Detailed studies, such as that by ArnaudFassetta and Provansal (1999) highlight how an intricate interplay of catchment changes drive delta evolution and growth. The largest river catchments feed the biggest and most rapidly expanding deltas (Figure 13.2). Where those catchments drain tectonically active hinterlands or highly erodible lithologies, the resulting deltas can be anomalously large. Thus the Akheloos delta, the largest on the Ionian Sea coast, drains soft flysch terrain in a seismically active region (Piper and Panagos 1981). Tectonically subsiding gulfs and shelves also favour delta development, with subsidence creating accommodation space and preservation conditions for deltaic sediments to accumulate. In contrast, excessively deep basins, frequent storms, or strong currents limit the extent of deltas, creating undersized deltas such as the Var and Ebro. The largest deltas in the Mediterranean are major depocentres containing some of the thickest Pliocene-Quaternary records in the region (Figure 13.2). The Po delta, for example, covers an area of approximately 770 km2 , has a Holocene record of 30 m in thickness and a total Plio-Quaternary sediment record of c.5,000 m. Regardless of the driving forces of late Holocene coastal accretion, and despite the concomitant loss of cultural heritage, the great delta plains provide the modern Mediterranean with extensive tracts of fertile farmland, prime real estate for urban sprawl, tourist infrastructure and airports, and some of the best wildlife habitats in Europe (Chapters 9 and 23). But today these important coastal environments are under threat, since

Coastal Geomorphology and Sea-Level Change

most of the Mediterranean deltas are no longer advancing, but are in retreat.

Humans as Agents of Coastal Change In modern times the amount of sediment reaching the Mediterranean via the Nile River had been drastically cut, most specifically with the construction of the High Dam at Aswan, and the problem has been compounded by uncontrolled coastal development (Frihy 2001). Historical studies of the Nile shoreline changes show how a delta front that had advanced steadily throughout the eighteenth century had, by the end of that century, gone into reverse. Thus on the Rosetta promontory (Figure 13.3) this retreat has accelerated from rates of 18 m yr−1 at the beginning of the nineteenth century to 230 m yr−1 in the 1990s (Fanos 1995; Ahmed 2002). The net shortfall in Nile sediment input has a knockon effect on the Israeli Mediterranean littoral where recent decades have witnessed the aeolianite coastal cliffs eroding back at rates of 20 cm yr−1 (Zviely and Klein 2004). Throughout the twentieth century these soft cliffs have contributed approximately 24 × 106 m3 of sediment to the budget of Israeli beaches, offsetting the loss of Nile sand. A further source of sand supply to this coast is similarly declining as reductions in agricultural and pastoral activity allow natural vegetation to recolonize and stabilize the dunefields of this coast (Tsoar and Blumberg 2002; Chapter 14). To the west, the Tunisian coastline is similarly retreating everywhere, except for some localized areas at the mouths of active wadis or in sheltered bays and spits (Oueslati 1995). Again the main culprit is marine erosion from human interventions on the coast, but sea-level rise is also a factor. Many ancient (Punic and Roman) port and harbour installations are submerged by a few tens of centimetres, and the widespread intrusion of seawater is favouring the extension of sebkhas (flat supratidal environments of sedimentation devoid of vegetation) and chotts (marginal parts of sebkhas colonized by salt-tolerant vegetation) and the transformation of formerly inhabited and cultivated fields into unproductive land. These environmental changes are most acute in the northern part of the Gulf of Gabès, where inundation is accentuated by active tectonic subsidence (Paskoff and Oueslati 1991). It is a similar story along many parts of the northern Mediterranean. Shipwrecks along the Gulf of Lions littoral track the progressive landward displacement of the beach barrier (Barusseau et al. 1996). The reduced supply of sediment to the Ebro delta coast in recent years

405

had led to a negative sediment budget, erosional retreat of the shoreline (Jiménez and Sánchez-Arcilla 1993), and coarsening of the beachface sediments through preferential winnowing of the finer fractions (Guillén and Palanques 1996). In this region, coastal erosion is threatening the existence of the great dunefields that form the seaward fringes of the Guadalquivir, Rhône, Ebro, Acheleous, and Goksu deltas, which all rely on unhampered sand supply for their survival. Where mobile dunefields have encroached on agricultural and urban areas they have commonly been stabilized by vegetation and woodland. In some places, however, overgrazing, trampling, and other human interference has destabilized them, leading to wind erosion problems. According to van der Meulen and Salman (1996) almost 75 per cent of the Mediterranean’s coastal dunes had, in the preceding thirty years, been damaged or destroyed, largely as a result of tourism. Along the Spanish and French Mediterranean shores, 75–80 per cent of the coastal dunes have succumbed to these pressures (Corre 1991), and they are becoming similarly stressed in Turkey, Greece, Israel, and Tunisia (Chapter 14). The gradual disappearance of the dunes means that sandy beaches are also vanishing at an increasing rate. Eroding delta fronts and stabilizing dune systems are not the only processes that are starving Mediterranean beaches of sediment. The prime reason is that many of the region’s high-discharge rivers are dammed and clogged by flood-control schemes (Chapters 8 and 11), so sands and gravels are no longer making it to the coast. Many of these dams are needed to underpin the dramatic rise in water demand from the increasing mass tourism around the shores of the Mediterranean, a region which sees its 130 million coastal population swelled seasonally (Chapter 21). To cater for such influxes, tourist developments such as marinas and promenade waterfronts have taken more and more control of the littoral zone. Along much of the Mediterranean’s shores, the longshore distribution of what little sediment reaches the coast is impeded and diverted into a new set of temporary sediment sinks. In all corners of the Mediterranean realm, tourist development is placing the littoral zone under unprecedented pressures (e.g. Frihy et al. 1996; Garcia and Servera 2003; Burak et al. 2004; Chapter 23). New coastal management practices have emerged to tackle the changing character of the Mediterranean’s soft shores. In the 1990s, Spain’s Costa del Sol authorities switched from stabilizing an eroding coastline with groyne fields to replenishing them artificially (Malvárez

406

Iain Stewart and Christophe Morhange

Garcia et al. 2002). Such beach nourishment schemes have had mixed success because their aim was to maintain sufficient beach width for recreational purposes rather than to recreate the exact position of the former shoreline or to reduce the impact of flooding (ibid.). Often the new man-made beaches are not robust enough to stand up to the winter storms that frequently buffet them. Furthermore, the offshore sediment is often far more coarse, so many beaches have become highly consolidated or excessively shelly, giving too abrasive a surface for recreational activities. Nevertheless, the beach nourishment approach promoted in Spain is now the preferred coastal adaptation in countries such as Turkey, while others, such as Greece, prefer ‘hard-engineered’ protection schemes (Hanson et al. 2002). In Italy, coastal retreat is being tackled by a mixed set of practices, mostly minor beach nourishment schemes but some major coastal engineering interventions. The largest of these are the construction of a barrier beach in the Venetian Lagoon and a barrier beach front to the city itself, and similar large barrier beach projects at Ravenna in the Po River delta (Hanson et al. 2002). Along this Adriatic coast, the problem is not just disappearing beaches but the threat of imminent flooding. Although the Mediterranean is generally characterized by minor tidal and surge variability, more extreme flooding can characterize semi-enclosed gulfs and restricted marine settings, such as the north Aegean sea (Tsimplis and Blackman 1997). The problem is most acute, however, in the low-lying and tectonically subsiding northern shores of Adriatic Italy (Nicholls and Hoozemans 1996). Here, the city of Venice has long been flooded by storm surges, but the frequency of surges in the Venetian Lagoon has dramatically increased since the 1960s, reaching about two flooding events per year, the highest value since AD 792 (Camuffo and Sturaro 2004). The change reflects a relative sea-level rise of 0.3 m resulting from land subsidence due to a combination of onshore groundwater withdrawal and offshore methane extraction (e.g. Brunetti et al. 1998). Much of this subsidence appears to have ameliorated in recent years (Tosi et al. 2002), but the city of Venice remains at risk (for a review, see Fletcher and Spencer 2005). The reason for this continuing threat is accelerated global sea-level rise (Pirazzoli 2006). Although twentieth-century sea-level-rise trends are arguably overestimated (Cabannes et al. 2001) or underestimated (Douglas and Peltier 2002), global studies continue to indicate that world sea-level rise is accelerating (Church and White 2006). Satellite and tide-gauge data indicate that Mediterranean waters are currently on the up, with satellite altimetry data

showing that between 1992 and 1996 the mean rate of sea-level rise over the Mediterranean Sea was 7 ±1.5 mm yr−1 (Cazenave et al. 2002). The sea-level changes were spatially variable (e.g. 20–30 mm yr−1 in the Levantine basin and falling sea levels in the Ionian Sea), indicating that the sea-level rise was the result of thermal expansion—the differential heating of surface water layers. During roughly the same period (1994–7), a rapid rise of western Mediterranean sea level by more than 10 mm year−1 has been at the expense of a 40 per cent fall in the sea-level drop between the Atlantic and the Mediterranean from 1994 to 1997. This highlights the fluctuating exchange flow through the Strait of Gibraltar which in turn is driven indirectly by changing conditions in the Mediterranean (Tetjana et al. 2000). Given uncertainties concerning future warming scenarios and ocean response, a global rise in sea level ranging from about 0.2–0.9 m by the year 2100 appears possible, with the estimates being current towards the upper end of this range. Under such higher water-level conditions, low-lying portions of the Mediterranean coastline face heightened flooding threats in the coming century. Quantifying the extent of its vulnerability is difficult, not least because of the restless character of its land movements (Nicholls and Hoozemans 1996). Nevertheless, microtidal coasts such as the Mediterranean are especially vulnerable to rising water levels because the elevational growth range of coastal vegetation is tied to tidal range (Day et al. 1995). These authors also argue that sediment deficits already observed along soft deltaic shores are likely to become more acute with an acceleration in sea-level rise, because structural adaptations to protect against more frequent river floods and storm surges will further decouple beaches from their sediment sources. At 1990 figures, ten million people lived within the region’s 1-in-1,000-year storm surge level, and about one million people experienced coastal flooding in a typical year. But, if sea-level rise continues at the modern pace of 1.5–2 mm yr−1 witnessed over the last century (Miller and Douglas 2004), ever more people fall within the flood-risk zone. With the inundation of tectonically subsiding or stable coastal lowlands comes the concomitant loss of land, biodiversity, and fisheries (Nicholls and Hoozemans 1996). Nicholls et al. (1999), for example, warn that Mediterranean coastal wetlands could disappear almost completely by the 2080s due to sea-level rise alone (Chapter 9). At risk are the great delta plains that once harboured the coastal dwellers in Palaeolithic and Mesolithic times, and which today form much of the Mediterranean’s industrial, agricultural, and tourist heartland.

Coastal Geomorphology and Sea-Level Change

Conclusions The impacts of present and future climate change on the Mediterranean coastal zone are difficult to gauge because, in addition to the inherent uncertainties in eustatic change estimations, this fragmented and heterogeneous mosaic of marginal seas ensures a complex response to tectonic, climatic, and eustatic forcing mechanisms. To address this important societal issue requires more studies that refine our understanding of the tectonic and glacio-hydro-isostatic dynamics of its seaboard and the sedimentological and geomorphological dynamics of its shoreline. Such research efforts increasingly demand interdisciplinary collaborations between geologists, geographers, oceanographers, climate scientists, archaeologists, biologists, and a suite of other practitioners, a need perfectly exemplified by this volume.

Acknowledgements We thank A. Cundy, R. Gehrels, J. Laborel, P. Pirazzoli, and N. Marriner for helpful suggestions and comments, and Heather Viles and Jamie Woodward for critical and constructive reviews.

References Ahmed, M. H. (2002), Multi-temporal conflict of the Nile Delta coastal changes, Egypt, in Littoral 2002, The Changing Coast. EuroCoast, Portugal. Aksu, A. E., Mudie, P. J., Rochon, A., Kaminski, M., Abrajano, T., and Ya¸sar, D. (2002), Persistent Holocene outflow from the Black Sea to the eastern Mediterranean contradicts the Noah’s Flood hypothesis. GSA Today, 1–9 May. Altınok, Y. and Ersoy, S¸ . (2000), Tsunamis observed on and near the Turkish Coast. Natural Hazards 21: 185–205. Ambraseys, N. N. (1962), Data for the investigation of seismic sea waves in the eastern Mediterranean. Bulletin of the Seismological Society of America 52: 895–913. Antonioli, F., Chemello, R., Improta, S., and Riggio, S. (1999), Dendropoma lower intertidal reef formations and their palaeoclimatological significance, NW Sicily. Marine Geology 161: 155–70. Bard, E., Potter, E.-K., Silenzi, S., and Improta, S. (2004), 215-ka History of sea-level oscillations from marine and continental layers in Argentarola Cave speleothems. Global and Planetary Change 43: 57–78. Kershaw, S., Renda, P., Rust, D., Belluomini, G., Cerasoli, M., Radtke, U., and Silenzi, S. (2006), Elevation of the last interglacial highstand in Sicily (Italy): A benchmark of coastal tectonics. Quaternary International 145–6: 3–18. Armijo, R., Meyer, B., King, G. C. P., Rigo, A., and Papanastassiou, D. (1996), Quaternary evolution of the Gulf of Corinth rift and its implications for the Late Cenozoic evolution of the Aegean. Geophysical Journal International 126: 11–53.

407

Arnaud-Fassetta, G. and Provansal, M. (1999), High frequency variations of water flux and sediment discharge during the Little Ice Age (1586–1725 AD) in the Rhone Delta (Mediterranean France). Relationship to catchment. Hydrobiologia 410: 241–50. Bard, E., Hamelin, B., Arnold, M., Montaggioni, L., Cabioch, G., Faure G., and Rougerie, F. (1996), Deglacial sea-level record from Tahiti corals and the timing of global meltwater discharge. Nature 382: 241–4. Antonioli, F., and Silenzi, S. (2002), Sea level during the penultimate interglacial period based on submerged stalagmite from Argentarola Cave (Italy). Earth and Planetary Science Letters 196: 135–46. Barusseau, J. P., Akouango, É., Bâ, M., Deschamps, C., and Golf, A. (1996), Evidence for short-term retreat of the barrier shorelines. Quaternary Science Reviews 15: 763–71. Ben Gai, Y., Ben-Avraham, Z., Buchbinder, B., and Kendall, C. G. St. C. (2005), Post-Messinian evolution of the Southeastern Levant Basin based on two-dimensional stratigraphic simulation. Marine Geology 221: 359–71. Blackman, D. J. (2005), Archaeological evidence for sea level changes. Zeitschrift für Geomorphologie 137: 61–70. Blanc, P. L. (2002), The opening of the Plio-Quaternary Gibraltar Strait: assessing the size of the cataclysm. Geodinamica Acta 15: 303–17. Blanchon, P. and Shaw, J. (1995), Reef drowning during the last deglaciation: evidence for catastrophic sea-level rise and icesheet collapse. Geology 23: 4–8. Bordoni, P. and Valensise, G. (1998), Deformation of the 125 ka marine terrace in Italy: tectonic implications, in I. S. Stewart and C. Vita-Finzi (eds.), Coastal Tectonics. Geological Society, London, Special Publication 146: 71–110. Boudouresque, C. F. (1971), Contribution à l’étude phytosociologique des peuplements algaux des côtes varoises. Vegetatio 22: 83–184. Brückner, H. (2002), Holocene landscape evolution of Buyuk Menderes. Zeitschrift für Geomorphologie 127: 47–65. Müllenhoff, M., van der Borg, K., and Vött, A. (2004), Holocene coastal evolution of western Anatolia—the interplay between natural factors and human impact, in Human Records of Recent Geological Evolution in the Mediterranean Basin— Historical and Archaeological Evidence. CIESM Workshop Monograph 24: 51–6. Vött, A., Schriever, M., and Handl, M. (2005), Holocene delta progradation in the eastern Mediteraanean—case studies in their historical context. Méditerranée 1/2: 95–106. Brunetti, A., Denèfle, M., Fontugne, M., Hatté, C., and Pirazzoli, P. A. (1998), Sea-level and subsidence data from a Late Holocene back-barrier lagoon (Valle Standiana, Ravenna, Italy). Marine Geology 150: 29–37. Burak, S., Do˘gan, E., and Gazio˘glu, C. (2004), Impact of urbanisation and tourism on coastal environment. Ocean and Coastal Management 47: 515–27. Cabannes, C., Cazenave, A., and Le Provost, C. (2001), Sea level rise during past 40 years determined from satellite and in situ observations. Science 294: 840–2. Camuffo, D. and Sturaro, G. (2004), Use of proxy-documentary and instrumental data to assess the risk factors leading to sea flooding in Venice. Global and Planetary Change 40: 93–103. Carobene, L. and Dai Pra, G. (2003), Middle and Upper Pleistocene sea-level highstands along the Tyrrehenian coast of Basilicata (southern Italy). Il Quaternario 4: 173–202.

408

Iain Stewart and Christophe Morhange

Carter, R. W. G. (1988), Coastal Environments. Academic Press, New York. Cazenave, A., Bonnefond, P., Mercier, F., Dominh, K., and Toumazou, V. (2002), Sea level variations in the Mediterranean Sea and Black Sea from satellite altimetry and tide guages. Global and Planetary Change 34: 59–86. Church, J. A. and White, N. J. (2006), A 20th century acceleration in global sea-level rise. Geophysical Research Letters 33. Clauzon, G., Suc, J.-P., Gautier, F., Berger, A., and Loutre, M.F. (1996), Alternative interpretation of the Messinian salinity crisis: controversy resolved? Geology 24: 363–6. Collier, R. E. L., Leeder, M. R., and Jackson, J. A. (1995), Quaternary drainage development, sediment fluxes and extensional tectonics in Greece, in J. Lewin, M. G. Macklin, and J. C. Woodward (eds.), Mediterranean Quaternary River Environments. Balkema, Rotterdam, 31–44. Collina-Girard, J. (1997), Analyse de profils sous-marins levés en plongée sur les côtes provençales, implications eustatiques et néo-tectoniques. CR Acad. Sc. Paris 325: 955–9. (1998), Paliers bathymétriques observés en plongée au littoral occidental de la Corse. Conséquences néotectoniques. CR Acad. Sci. Paris 327: 121–6. (2001), L’Atlantide devant le détroit de Gibraltar? Mythe et géologie. CR Acad. Sci. Paris, Earth and Planetary Sciences 333: 233–40. (2002), La Crise Finiglaciaire à Gibraltar et l’Atlantide: Tradition orale et géologie. Préhistoire Anthroplogie Méditerranéennes 10/11: 53–60. Corre, J.-J. (1991), The sand dunes and their vegetation along the Mediterranean coast of France. Their likely response to climatic change. Landscape Ecology 6: 65–75. Cundy, A. (2005), Recent rapid sea-level change in the eastern Mediterranean and the coastal sedimentary record. Zeitschrift für Geomorphologie, Suppl.137: 29–35. Kortekaas, S., Dewez, T., Stewart, I. S., Collins, P. E. F., Croudace, I. W., Maroukian, H., Papanastassiou, D., GakiPapanastassiou, P., Pavlopoulos, K., and Dawson, A. (2000), Coastal wetlands as recorders of earthquake subsidence in the Aegean: a case study of the 1894 Gulf of Atalanti earthquakes, central Greece. Marine Geology 170: 3–26. Davis Jr., R. A. (1996), Coasts. Prentice Hall, Upper Saddle River, NJ. Dawson, A. G. (1996), The geological significance of tsunamis. Zeitschrift für Geomorphologie 102: 199–210. and Stewart, I. S. (2007), Tsunami deposits in the geological record. Sedimentary Geology 200: 166–83. Day, J. W., Pont, D., Hensel, P. F., and Ibanez, C. (1995), Impacts of sea-level rise on deltas in the Gulf of Mexico and the Mediterranean: the importance of pulsing events to sustainability. Estuaries 18: 636–47. De Guidi, C., Catalano, S., Monaco, C., and Tortoni, L., (2003), Morphological evidence of Holocene coseismic deformation in the Taormina region (NE Sicily). Journal of Geodynamics 36: 193–211. Di Vita, A. (1996), Earthquakes and civil strife at Gortyn (Crete) in the period between Justinian and Constant II (6–7th century AD), in S. Stiros and R. E. Jones (eds.), Archaeoseismology, British School at Athens Fitch Laboratory Occasional Paper 7: 45–50. Dominey-Howes, D. T. M., Cundy, A. B., and Croudace, I. W. (2000), High energy marine flood deposits on Astypalaea

Island, Greece: Possible evidence for the AD 1956 southern Aegean tsunami. Marine Geology 163: 303–15. Dawson, A., and Smith, D. (1998), Late Holocene coastal tectonics at Falasarna, W. Crete: a sedimentary study, in I. S. Stewart and C. Vita-Finzi (eds.), Coastal Tectonics, Geological Society, London, Special Publication 146: 343–52. Douglas B. C. and Peltier, W. R. (2002), The Puzzle of Global SeaLevel Rise. Physics Today 55: 35–40. Dubar, M. and Anthony, E. J. (1995), Holocene environmental change and river-mouth sedimentation in the Baie des Anges, French Riviera. Quaternary Research 43: 329–43. Duggen, S., Hoernle, K., van den Bogaard, P., Rüpke, L., and Philips Morgan, J. C. (2003), Deep roots for the Messinian salinity crisis. Nature 422: 602–6. and Harris, C. (2004), Magmatic evolution of the Alboran region: The role of subduction in forming the western Mediterranean and causing the Messinian Salinity Crisis. Earth and Planetary Science Letters 218: 91–108. Dumas, B., Gueremy, P., Lhenaff, R., and Raffy, J. (1993), Rapid uplift, stepped marine terraces and raised shorelines on the Calabrian coast of the Messina Straits, Italy. Earth Surface Processes and Landforms 18: 241–56. Dvorak, J. J. and Mastrolorenzo, G. (1991), The mechanisms of recent vertical crustal movements in Campi Flegrei caldera, southern Italy. Special Paper of the Geological Society of America 263: 1–47. Emery, K. O., Aubrey, D. G., and Goldsmith, V. (1988), Coastal neo-tectonics of the Mediterranean from tide-gauge records. Marine Geology 81: 41–52. Fabbri, P. (1985), Coastline variations in the Po delta since 2500 BP. Zeitschrift für Geomorphologie 57: 155–67. Fairbanks R. G. (1989), A 17,000-year glacio-eustatic sea level record: influence of glacial melting rates on the Younger Dryas event and deep-ocean circulation. Nature 342: 637–2. Faivre, S. and Fouache, E. (2003), Some tectonic influences on the Croatian shoreline evolution in the last 2000 years. Zeitschrift für Geomorphologie 47: 521–37. Fanos, A. M. (1995), The impact of human activities on the erosion and accretion of the Nile Delta coast. Journal of Coastal Research 11: 821–33. Ferranti, L., Antonioli, F., Mauz, B., Amorosi, A., Dai Pra, G., Mastronuzzi, G., Monaco, C., Orrù, P., Pappalardo, M., and Radtke, U. (2006), Markers of the last interglacial sea-level high stand along the coast of Italy: Tectonic implications. Quaternary International 145/6: 30–54. Firth, C., Stewart, I., McGuire, W. J., Kershaw, S., and Vita-Finzi, C. (1996), Coastal elevation changes in eastern Sicily; implications for volcano instability, in W. J. McGuire, A. P. Jones, and J. Neuberg (eds.), Volcano Instability on the Earth and Other Planets. Geological Society, London, Special Publication 110: 153–67. Flemming, N. C. (1969), Archaeological evidence for eustatic change in sea level and earth movements in the Western Mediterranean in the last 2000 years. Geological Society of America, Special Paper 109. (1993), Predictions of relative coastal sea-level change in the Mediterranean based on archaeological, historical and tide-gauge data in L. Jeftic, J. D. Milliman, and G. Sestini (eds.), Climate Change and the Mediterranean. Edward Arnold, London, 247–81.

Coastal Geomorphology and Sea-Level Change (1978), Holocene eustatic changes and coastal tectonics in the north-east Mediterranean: implications for models of crustal consumption. Philosophical Transactions of the Royal Society of London A289: 405–58. (1998), Archaeological evidence for vertical movement on the continental shelf during the Palaeolithic, Neolithic and Bronze Age periods, in I. S. Stewart and C. Vita-Finzi (eds.), Coastal Tectonics. Geological Society, London, Special Publication 146: 129–46. Fletcher, C. A. and Spencer, T. (2005), Flooding and Environmental Challenges for Venice and its Lagoon. Cambridge University Press, Cambridge. Fornós, J. J., Gelabert, B., Ginés, À., Ginés, J., Tuccimei, P., and Vesica, P. (2002), Phreatic overgrowths on speleothems: a useful tool in structural geology in littoral karstic landscapes. The example of eastern Mallorca (Balearic Islands). Geodinamica Acta 15: 113–25. Fouache, É., Dalongeville, R., Kunesch, S., Suc, J.-P., Subally, D., Prieur, A., and Lozouet, P. (2005), The environmental setting of the harbor of the Classical sites of Oeniades on the Acheloos delta, Greece. Geoarchaeology 20: 285–302. Desruelles, S., Pavlopoulos, K., Dalgonville, R., Coquinot, Y., Peulvast, J.-P., and Potdevin, J.-L. (2005), Using beachrocks as sea level indicators in the insular group of Mykonos, Delos and Rhenia (Cyclades, Greece). Zeitschrift für Geomorphologie 137: 37–43. Frihy, O. E. (2001), The necessity of environmental impact assessment (EIA) in implementing coastal projects: lessons learned from the Egyptian Mediterranean coast. Ocean and Coastal Management 44: 489–516. Dewidar, K. M., and El Raey, M. M. (1996), Evaluation of coastal problems at Alexandria. Ocean and Coastal Management 30: 281–95. Galili, E. and Sharvit, J. (1998), Ancient coastal installations and the tectonic stability of the Israeli coast in historical times, in I. S. Stewart and C. Vita-Finzi (eds.), Coastal Tectonics. Geological Society, London, Special Publication 146: 147–63. Gamble, C. (1986), The Palaeolithic Settlement of Europe. Cambridge University Press, Cambridge. Garcia, C. and Servera, J. (2003), Impacts of tourist development on water demand and beach degradation on the island of Mallorca (Spain). Geografiska Annaler 85A: 287–300. Görür, N., Ca˘gatay, M. N., Emre, Ö., Alpar, B., Sakinç, M., Ã˙Islamo˘glu, Y., Algan, O., Erkal, T., Keçer, M., Akkök, and Karlık, G. (2001), Is the abrupt drowning of the Black Sea shelf at 7150 yr BP a myth? Marine Geology 176: 65–73. Got, H., Aloisi, J.-C., Monaco., A. (1985), Sedimentary processes in Mediterranean deltas and shelves, in S. J. Stanley and F.-C. Wetzel (eds.), Geological Evolution of the Mediterranean Basin. Springer, New York, 355–76. Goy, J. L. and Zazo, C. (1988), Sequences of Quaternary marine levels in Elche Basin (eastern Betic Cordillera, Spain). Palaegeography, Palaeoclimatology, Palaeoecology 68: 301–10. Zazo, C., and Dabrio, C. J. (2003), A beach-ridge progradation complex reflecting periodical sea-level and climate variability during the Holocene (Gulf of Almería, Western Mediterranean). Geomorphology 50: 251–68. Grenon, M. and Batisse, M. (eds.) (1989), Futures for the Mediterranean Basin: The Blue Plan. Oxford University Press, Oxford. Grove, A. T. and Rackman, O. (2001), The Nature of Mediterranean Europe: An Ecological History. London, Yale University Press.

409

Guillén, J. and Palanques, A. (1996), Short- and medium-term grain size changes in deltaic beaches (Ebro Delta, NW Mediterranean). Sedimentary Geology 101: 55–67. Hanson, H., Brampton, A., Capobianco, M., Dette, H. H., Hamm, L., Laustrup, C., Lechuga, A., and Spanhoff, R. (2002), Beach nourishment projects, practices and objectives—a European overview. Coastal Engineering 47: 81–111. Haslett, S. K. (2000), Coastal Systems. Routledge, London. Hodell, D. A., Curtis, J. H., Siero, F. J., and Raymo, M. E. (2001), Correlation of late Miocene to early Pliocene sequences between the Mediterranean and North Atlantic. Palaeooceanography 16: 164–78. Hsü, K. J., Ryan, W. B. F., and Cita, M. B. (1973). Late Miocene dessication of the Mediterranean. Nature 242: 240–4. Inman, D. L. (1994), Types of coastal zones, in Environmental Science in the Coastal Zone. Washington, DC, National Research Council, 67–84. and Jenkins, S. A. (1984), The Nile littoral cell and man’s impact on the coastal zone of the southeastern Mediterranean, in Proceedings 19th Coastal Engineering Conference, American Society of Civil Engineers, 2: 1600–17. and Nordstrom, C. E. (1971), On the tectonic and morphologic classification of coasts. Journal of Geology 79: 1–21. Jacobsen, T. W. (1976), 17,000 years of Greek prehistory. Scientific American 234: 76–87. Jiménez, J. A. and Sánchez-Arcilla, A. (1993), Medium-term coastal response at the Ebro delta, Spain. Marine Geology 114: 105–18. Kelletat, D. and Schellman, G. (2002), Tsunami in Cyprus: field evidences and 14 C dating results. Zeitschrift für Geomorphologie, Suppl. 46: 19–34. Kowalczyk, G., Schröder, B., and Winter, K. P (1976), A synoptic view on the neotectonic development of the Peloponnesian coastal regions. Zeitschrift der Deutschen Gesellschaft für Geowissenschaften 127: 447–65. Keraudren, B. and Sorel, D. (1987), The terraces of Corinth (Greece)—a detailed record of eustatic sea-level variations during the last 500,000 years. Marine Geology 77: 99–107. Kerey, I. E., Meriç, E., Tuno˘glu, C., Kelling, G., Brenner, R. L., and Do˘gan, A. U. (2004), Black Sea–Marmara Sea Quaternary connections: new data from the Bosphorus, Istanbul, Turkey. Palaeogeography, Palaeoclimatology, Palaeoecology 204: 277–95. Kershaw, S. (2000), Quaternary reefs of northeastern Sicily: Structure and growth controls in an unstable tectonic setting. Journal of Coastal Research 16/4: 1037–62. and Guo, L. (2001), Marine notches in coastal cliffs; indicators of relative sea-level change, Perachora peninsula, central Greece. Marine Geology 179: 213–28. Kraft, J. C., Aschenbrenner, S. E., and Rapp Jr., G., (1977), Application of Holocene stratigraphy to paleogeographic reconstructions of coastal Aegean archaeological sites. Science 195: 941–7. Rapp, G., Kayan, I., and Luce, J. V. (2003), Harbor areas at ancient Troy: sedimentology and geomorphology complement Homer’s Iliad. Geology 31: 163–6. Krijgsman, W. (2002), The Mediterranean: Mare Nostrum of Earth Sciences. Earth and Planetary Science Letters 205: 1–12. Hilgen, F. J., Raffi, I., Siero, F. J., and Wilson, D. S. (1999), Chronology, causes and progression of the Messinian salinity crisis. Nature 400: 652–5.

410

Iain Stewart and Christophe Morhange

Laborel, J. (1979), Fixed marine organisms as biological indicators for the study of recent sea level and climatic variations along the Brazilian tropical coast. Proc. 1978 International Symposium on Coastal Evolution in the Quaternary. São Paulo, 193–211. (1986) Vermetid gastropods as sea level indicators, in O. van de Plaasche (ed.), Sea Level Research, a Manual for Collection and Evaluation of Data. Geo Books, Norwich, 28–310. and Laborel-Deguen, F. (1994), Biological indicators of relative sea-level variation and of co-seismic displacements in the Mediterranean area. Journal of Coastal Research 10: 395–415. (1996), Biological indicators of Holocene sea-level and climatic variations on rocky coasts of tropical and subtropical regions, Quaternary International 31: 53–60. Morhange, C., Lafont, R., Le Campion, J., Laborel-Deguen, F., and Sartoretto, S. (1994), Biological evidence of sea-level rise during the last 4500 years on the rocky coasts of continental southwestern France and Corsica. Marine Geology 120: 203–23. Morhange, C., Collina-Girard, J., and Laborel-Deguen, F. (1999), Littoral bio-erosion, a tool for the study of sea level variations during the Holocene. Bulletin of the Geological Society of Denmark 45: 144–8. Lambeck, K. (1995), Late Pleistocene and Holocene sea-level change in Greece and south-western Turkey: a separation of eustatic, isostatic and tectonic contributions. Geophysical Journal International 122: 1022–44. (1996), Sea-level change and shore-line evolution in Aegean Greece since Upper Palaeolithic time. Antiquity 70: 588–611. and Bard, E. (2000), Sea-level change along the French Mediterranean coast since the time of the Last Glacial Maximum. Earth and Planetary Science Letters 175: 203–22. and Chappell, J. (2001), Sea level change through the last glacial cycle. Science 292: 679–86 and Purcell, A. (2005), Sea-level change in the Mediterranean since the LGM: Model predictions for tectonically stable areas. Quaternary Science Reviews 24: 1969–88. Antonioli, F., Purcell, A., and Silenzi, S. (2004a), Sea-level change along the Italian coast for the past 10,000 yrs. Quaternary Science Reviews 23: 1567–98. Anzidei, M., Antonioli, F., Benini, A., and Esposito, A. (2004b), Sea level in Roman times in the Central Mediterranean and implications for recent change. Earth and Planetary Science Letters 224: 563–75. Leeder, M. R., Seger, M., and Stark, C. P. (1991), Sedimentology and tectonic geomorphology adjacent to active and inactive normal faults in the Megara Bain and Alkonides Gulf, Central Greece. Journal of the Geological Society, London 148: 331–43. Loget, N., van den Driessche, J., and Davy, P. (2005), How did the Messinian Salinity Crisis end? Terra Nova 17: 414–18. Major, C., Ryan, W., Lericolais, G., and Hajdas, I. (2002), Constraints on Black Sea outflow to the Sea of Marmara during the last glacial–interglacial transition. Marine Geology 190: 19–34. Malvárez Garcia, G., Pollard, J., and Domínguez Rodríguez, R. (2000), Origins, management and measurement of stress on the coast of southern Spain. Coastal Management 28: 215–34. Hughes, R. (2002), Coastal zone management on the Costa del Sol: a small business perspective. Journal of Coastal Research 36: 470–82.

Marriner, N. and Morhange, C. (2006), The ‘Ancient Harbour Parasequence’: Anthropogenic forcing of the stratigraphic highstand record. Sedimentary Geology 186: 13–17. (2007), Geoscience of ancient Mediterranean harbours. Earth-Science Reviews 80: 137–94. Morhange, C., Doumet-Serhal, C., and Carbonel, P. (2006), Geoscience discovers Phoenicia’s buried harbors. Geology 34: 1–4. Mastronuzzi, G. and Sansò, P. (2000), Boulder transport by catastrophic waves along the Ionian coast of Apulia (southern Italy). Marine Geology 170: 93–103. (2004), Large boulder accumulations by extreme waves along the Adriatic coast of southern Apulia (Italy). Quaternary International 120: 173–84. Miller, L. and Douglas, B. C. (2004), Mass and volume contributions to twentieth- century global sea level rise. Nature 428: 406–9. Milliman, J. D. (1992), Sea-level response to climate change and tectonics in the Mediterranean Sea, in L. Jeftic (ed.), Climatic Change and the Mediterranean. Edward Arnold, London, 45–57. Miyauchi, J., Dai Pra, G., and Sylos Labini, S. (1994), Geochronology of Pleistocene marine terraces and regional tectonics in the Tyrrhenian coast of South Calabria, Italy. Il Quaternario 7: 17–34. Monaco, C., Scicchitano, G., and Tortorici, L. (2006), Large boulder accumulations by tsunami waves along the south-eastern coast of Sicily. Geophysical Research Abstracts 8: 02643. Morhange, C., Laborel, J., and Hesnard, A. (2001), Changes of relative sea level during the past 5000 years in the ancient harbour of Marseilles, Southern France. Palaeogeography, Palaeoclimatology, Palaeooecology 166: 319–29. Blanc, F., Bourcier, M., Carbonel, P., Prone, A., SchmittMercury, S., Vivent, D., and Hesnard, A. (2003), Biosedimentology of the late Holocene deposits of the ancient harbor of Marseilles (Southern France, Mediterranean sea). The Holocene 13/4: 593–604. Marriner, N., Laborel, L., Todesco, M., and Oberlin, C. (2006), Rapid sea-level movements and noneruptive crustal deformations in the Phlegrean Fields caldera, Italy. Geology 34: 93–6. Mörner, N.-A. (1996), Sea level variability. Zeitschrift für Geomorphologie, NS 102: 223–32. Murray, S., Coleman, J., Roberts, H. H., and Salama, M. (1981), Accelerated currents and sediment transport off the Damietta Nile promontory. Nature 292: 51–4. Naylor, L. and Viles, H. A. (2002), A new technique for evaluating short-term rates of bioerosion and bioprotection. Geomorphology 47: 31–44. Nicholls, R. J. and Hoozemans, F. M. J. (1996), The Mediterranean: vulnerability to coastal implications of climate change. Ocean and Coastal Management 31: 105–32. and Marchand, M. (1999), Increasing flood risk and wetland losses due to global sea-level rise: regional and global analyses. Global Environmental Change 9: S69– S87. Orsi, G., Civetta, L., Del Gaudio, C., de Vita, S., Di Vito, M. A., Isaia, R., Petrazzuoli, S. M., Ricciardi, G. P., and Ricco, C. (1999), Short-term ground deformations and seismicity in the resurgent Camp Flegrei caldera (Italy): an example of active blockresurgence in a densely populated area. Journal of Volcanology and Geothermal Research 91: 415–51.

Coastal Geomorphology and Sea-Level Change Oueslati, A. (1995), The evolution of low Tunisian coasts in historical times: from progradation to erosion and salinisation. Quaternary International 29/30: 41–7. Papadopoulos, G. A. (1998), A Tsunami Catalogue of Greece and Adjacent Seas. EC Project GITEC-TWO, Institute of Geodynamics, National Observatory of Athens, 8. Paskoff, R. and Oueslati, A. (1991), Modifications of coastal conditions in the Gulf of Gabès (Southern Tunisia) since Classical Antiquity. Zeitschrift für Geomorphologie, Suppl. 81: 149–62. Peltier, W. R. (1987), Mechanisms of relative sea-level change and the geophysical responses to ice-water loading, in R. J. N. Devoy (ed.), Sea Surface Studies. Croom Helm, London, 57–94. (1998), Global glacial adjustment and coastal tectonics, in I. S. Stewart and C. Vita-Finzi (eds.), Coastal Tectonics. Geological Society, London, Special Publication 146: 1–30. (2000), On eustatic sea level history: Last Glacial Maximum to Holocene. Quaternary Science Reviews 21: 377–96. (2001), Global slacial isostatic adjustment and modern instrumental records of relative sea-level history, in B. Douglas, M. S. Kearney, and S. P. Leatherman (eds.), Sea Level Rise: History and Consequences. Academic Press, San Diego, 65–96. (2004), Global glacial isostatic adjustment and the surface of the ice-age Earth: The ICE-5G(VM2) model and GRACE. Annual Reviews of Earth and Planetary Sciences 32: 111–49. Péres, J. M. (1982), Zooations and organismic assemblages, in O. Kinne (ed.), Marine Ecology: A Comprehensive, Integrated Treatise on Life in Oceans and Coastal Waters. Wiley, New York, i. Petit Maire, N. (2004), Climatic crises and man in the Mediterranean basin: the last 20,000 years, in Human Records of Recent Geological Evolution in the Mediterranean Basin—Historical and Archaeological Evidence. CIESM Workshop monograph 24: 17–24. Piper, D. J. W. and Panagos, A. G. (1981), Growth patterns of the Acheloos and Evinos deltas, western Greece. Sedimentary Geology 28: 111–32. Pirazzoli, P. 1976. Sea level variations in the Northwest Mediterranean during Roman times, Science, 194, 519–521. (1986a), Marine notches, in O. van de Plaasche (ed.), Sealevel Research: A Manual for the Collection and Evaluation of Data. Geo Books, Norwich, 361–400. (1986b), The Early Byzantine Tectonic Paroxysm. Zeitschrift für Geomorphologie, NS 62: 31–49. (1991), World Atlas of Holocene Sea-Level Changes, Oceanography Series. Elsevier, Amsterdam. (1998), A comparison between postglacial isostatic predictions and late Holocene sea-level field data from Mediterranean and Iranian coastal areas. GeoResearch Forum 3/4: 401–20. (2005), A review of possible eustatic, isostatic and tectonic contributions in eight late-Holocene relative sea-level histories from the Mediterranean area. Quaternary Science Reviews 24: 1989–2001. (2006), The projected ‘MOSE’ barriers against flooding in Venice (Italy) and the expected global sea-level rise. Journal of Marine Environmental Engineering 8: 247–61. Ausseil-Badie, J., Giresse, P., Hadjidaki, E., and Arnold, M. (1992), Historical environmental changes at

411

Phalasarna harbour, West Crete. Geoarchaeology 7: 373– 92. Stiros, S. C., Arnold, M., Laborel, J., Laborel-Deguen, F., and Papageorgiou, S. (1994a), Episodic uplift deduced from Holocene shorelines in the Perachora Peninsula, Corinth area, Greece. Tectonophysics 229: 201–9. Laborel, J., Laborel-Deguen, F., Arnold, M., Papageorgiou, S., and Morhange, C. (1994b), Late Holocene shoreline changes related to palaeoseismic events in the Ionian Islands, Greece. The Holocene 4: 397–405. Laborel, J., and Stiros, S. C. (1996a), Coastal indicators of rapid uplift and subsidence: examples from Crete and other Eastern Mediterranean sites, Zeitschrift für Geomorphologie, Suppl. NS 102: 21–35. (1996b), Earthquake clustering in the Eastern Mediterranean during historical times, Journal of Geophysical Research 101 B3: 6083–97. Saliege, J. F., Erol, O., Kayan, I., and Person, A. (1991), Holocene raised shorelines on the Hatay coast (Turkey) Paleoecological and tectonic implications, Marine Geology 96: 295–311. Price, S., Higham, T., Nixon, L., and Moody, J. (2002), Relative sea-level changes in Crete: Reassessment of radiocarbon dates from Sphakia and west Crete. Annual of the British School at Athens 97: 171–200. Rodríguez-Vidal, J., Cáceres, L. M., Finlayson, J. C., Garcia, F. J., and Martínez-Aguirre, A. (2004), Neotectonics and shoreline history of the Rock of Gibraltar, southern Iberia. Quaternary Science Reviews 23: 2017–29. Rust, D. and Kershaw, S. (2000), Holocene tectonic uplift patterns in northeastern Sicily: evidence from marine notches in coastal outcrops. Marine Geology 167: 105–26. Ryan, W. B. F. and Pitman, W. C. (2000), Noah’s Flood: The New Scientific Discoveries About the Event that Changed History. Touchstone, New York. Major, C. O., Shimkus, K., Moskalenko, V., Jones, G. A., Dimitrov, P., Görür, N., Sakinç, M., and Yüce, H. (1997), An abrupt drowning of the Black Sea Shelf. Marine Geology 138: 119–26. Major, C. O., Lericolais, G., and Goldstein, S. L. (2003), Catastrophic flooding of the Black Sea. Annual Reviews in Earth and Planetary Sciences 31: 525–54. Said, R. (1981), The Geological Evolution of the River Nile. Springer, New York. Siddall, M., Pratt, L. J., Helfrich, K. R., and Gosan, L. (2004), Testing the physical oceanographic implications of the suggested sudden Black Sea infill 8400 years ago. Palooceanography 19: doi:10.1029/2003PA000903. Sivan, D., Wdowinski, S., Lambeck, K., Galili, E., and Raban, A. (2004), Holocene sea-level changes along the Mediterranean coast of Israel, based on archaeological observations and numerical model. Palaeogeography, Palaeoclimatology, Palaeooecology 167: 101–17. Soloviev, S. L. (1990), Tsunamigenic zones in the Mediterranean Sea. Natural Hazards 3: 183–202. Solovieva, O. N., Go, C. N., Kim, K. S., and Shchetnikov, N. A. (2000), Tsunamis in the Mediterranean Sea, 2000 BC– 2000 AD. Kluwer Academic, Dordrecht. Soter, S. (1998), Holocene uplift and subsidence of the Helike Delta, Gulf of Corinth, Greece, in I. S. Stewart and C. VitaFinzi (eds.), Coastal Tectonics. Geological Society, London, Special Publication 146: 41–56.

412

Iain Stewart and Christophe Morhange

Stanley, D. J. (1989), Subsidence of the northeastern Nile Delta: Rapid rates, possible causes, and consequences. Science 240: 497–500. and Warne, A. G. (1994), Worldwide initiation of Holocene marine deltas: deceleration of sea-level rise as a principle factor. Science 265: 228–31. (1997), Holocene sea-level change and early human utilisation of deltas. GSA Today 7/12: 1–7. Goddio, F., Jorstad, T. F., and Schnepp, G. (2004), Submergence of ancient Greek cities off Egypt’s Nile Delta. GSA Today 14: 4–10. Stewart, I. (1996), Holocene uplift and palaeoseismicity on the Eliki Fault, western Gulf of Corinth, Greece. Annali di Geofisica 39: 575–88. Cundy, A., Kershaw, S., and Firth, C. (1997), Holocene coastal uplift in the Taormina area, northeastern Sicily: implications for the southern prolongation of the Calabrian seismogenic belt. Journal of Geodynamics 24: 37–50. Stiros, S. C. (2000), Fault pattern of Nisyros Island volcano (Aegean Sea): structural and archaeological evidence, in W. J. McGuire, D. R. Griffiths, P. L. Hancock, I. S. Stewart, and C. Vita-Finzi (eds.), The Archaeology of Geological Catastrophes. Geological Society, London, Special Publication 171: 385–97. and Papageorgiou, S. (2001), Seismicity of Western Crete and destruction of the town of Kisamos at AD 365: Archaeological evidence. Journal of Seismology 5: 381–97. Arnold, M., Pirazzoli, P. A., Laborel, J., Laborel, F., and Papageorgiou, S. (1992), Historical coseismic uplift on Euboea Island, Greece. Earth and Planetary Science Letters 108: 109–17. Suri´c, M., Juraˇci´c, M., Horvatiniˇci´c, N. and Broni´c, I. K. (2005), Late Pleistocene-Holocene sea-level rise and the pattern of coastal karst inundation: records from submerged speleothems along the Eastern Adriatic Coast (Croatia). Marine Geology 214: 163–75. Tetjana, R., Garrett, C., and Le Traon, P. Y. (2000), Western Mediterranean sea-level rise: Changing exchange flow through the Strait of Gibraltar. Geophysical Research Letters 27: 2949–52. Thommeret, Y., Thommeret, J., Laborel, J., Montaggioni, L. F., and Pirazzoli, P. A. (1981), Late Holocene shoreline changes and seismo-tectonic displacements in western Crete (Greece). Zeitschrift für Geomorphologie, NS 40: 127–49. Tinti, S. and Maramai, A. (1996), Catalogue of tsunamis generated in Italy and in Côte d’Azur, France: a step towards a unified catalogue of tsunamis in Europe. Annali di Geofisica 39: 1253–99. Tosi, L., Carbognin, L., Teatini, P., Strozzi, T., and Wegmüller, U. (2002), Evidence of the present relative land stability of Venice, Italy, from land, sea, and space observations. Geophysical Research Letters 29: doi:1029/2001GL013211. Trenhaile, A. S. (2002), Rock coasts, with particular emphasis on shore platforms. Geomorphology 48: 7–22. Tsimplis, M. N. and Blackman, D. (1997), Extreme sea-level distribution and return periods in the Aegean and Ionian Seas. Estuarine, Coastal and Shelf Science 44: 79–89. Tsoar, H. and Blumberg, D. G. (2002), Formation of parabolic dunes from barchan and transverse dunes along Israel’s Mediterranean coast. Earth Surface Processes and Landforms 27: 1147–61. Tuccimei, P., Fornós, J. J., Ginés, À., Ginés, J., Gràcia, F., and Mucedda, M. (2003), Sea level change at Capo Caccia

(Sardinia) and Malorca (Balearic Islands) during oxygen isotope substage 5e, based on Th/U datings of phreatic overgrowths on speleothems, in Puglia 2003—Final Conference Project IGCP 437. Ulu˘g, A., Duman, M., Ersoy, S¸ ., Özel, E., and Avcı, M. (2005), Late Quaternary sea-level change, sedimentation and neotectonics of the Gulf of Gökova, southeastern Aegean Sea. Marine Geology 221: 381–95. van Andel, T. H. (1989), Late Quaternary sea-level changes and archaeology. Antiquity 63: 733–46. Jacobsen, T. W., Jolly, J. B., and Lianos, N. (1980), Late Quaternary history of the coastal zone near Franchthi Cave, southern Argolid, Greece. Journal of Field Archaeology 7: 389–402. Zangger, E., and Perissoratis, C. (1990), Quaternary transgressive/regressive cycles in the Gulf of Argos, Greece. Quaternary Research 34: 317–29. van de Plaasche, O. (ed.) (1986), Sea-Level Research: A Manual for the Collection and Evaluation of Data. Geo Books, Norwich, xxi. van der Meulen, F. and Salman, A. H. P. M. (1996), Management of Mediterranean coastal dunes. Ocean and Coastal Management 30: 177–95. Vella, C. and Provansal, M. (2000), Relative sea-level rise and neotectonic events during the last 6500 yr on the southern eastern Rhône delta, France. Marine Geology 170: 27–39. Fleury, T.-J., Raccasi, G., Provansal, M., Sabatier, F., and Bourcier, M. (2005), Evolution of the Rhône delta plain in the Holocene. Marine Geology 222/3: 235–65. Vesica, P. L., Tuccimei, P., Turi, B., Fornós, J. J., Ginés, À., and Ginés, J. (2000), Late Pleistocene palaeoclimates and sea-level change in the Mediterranean as inferred from stable isotope and U-series studies of overgrowths on speleothems, Mallorca, Spain. Quaternary Science Reviews 19: 865–79. Warny, S. A., Bart, P. J., and Suc, J.-P. (2003), Timing and progression of climatic, tectonic and glacioeustatic influences on the Messinian Salinity Crisis. Palaegeography, Palaeoclimatology, Palaeoecology 202: 59–66. Westaway, R. (1993), Quaternary uplift of southern Italy. Journal of Geophysical Research 87: 21,741–72. Whelan, F. and Kelletat, D. (2002), Geomorphic evidence and relative and absolute results for tsunami events in Cyprus. Science of Tsunami Hazards 20: 3–18. Wood, P. B. (1996), Dating and Origin of Late Quaternary Catastrophic Shoreline Activity around the Mediterranean Sea. Unpublished Ph.D. thesis, Royal Holloway College, University of London. Wyatt, T. (2004), Geochronology and myth—are gods catastrophes? in Human Records of Recent Geological Evolution in the Mediterranean Basin—Historical and Archaeological Evidence. CIESM Workshop monograph 24: 119–30. Yaltirak, C., Sakinç, M., Aksu, A. E., Hiscott, R. N., Galleb, B., and Ulgen, U. B. (2002), Late Pleistocene uplift history along the southwestern Marmara Sea determined from raised coastal deposits and global sea-level variations. Marine Geology 190: 282–305. Yalçıner, A. C., Alpar, B., Altınok, Y., Özbay, ˙I., and Inamura, F. (2002), Tsunamis in the Sea of Marmara: Historical documents for the past, models for the future. Marine Geology 190: 445–63. Zazo, C., Silva, P. G., Goy, J. L., Hillaire-Marcel, C., Ghaleb, B., Lario, J., Bardají, T., and González, A. (1999), Coastal uplift in continental collision plate boundaries: data from the Last

Coastal Geomorphology and Sea-Level Change Interglacial marine terrace of the Gibraltar Straits area (south Spain). Tectonophysics 301: 95–109. Goy, J. L., Dabrio, C. J., Bardají, T., Hillaire-Marcel, C., Ghaleb, B., González-Delgado, J. Á., and Soler, V. (2003), Pleistocene raised marine terraces of the Spanish and Atlantic coasts: records of coastal uplift, sea-level highstands and climate changes. Marine Geology 194: 103–33.

413

Zerbini, S. (and 16 others) (1996), Sea level in the Mediterranean: a first step towards separating crustal movements and absolute sea-level variations. Global and Planetary Change 14: 1–48. Zviely, D. and Klein, M. (2004), Coastal cliff retreat rates at BeitYannay, Israel, in the 20th century. Earth Surface Processes and Landforms 29: 175–84.

This chapter should be cited as follows Stewart, I. S. and Morhange, C. (2009), Coastal geomorphology and sea-level change, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 385–413.

This page intentionally left blank

14

Aeolian Processes and Landforms Andrew Goudie

Introduction Certain parts of the Mediterranean lands are drylands— notably south-east Spain, the North African littoral, and parts of the Levant. This means that there is potential for aeolian processes to operate locally, especially where the vegetation cover has been depleted by human activities. Although water erosion is probably the most pervasive cause of land degradation in the Mediterranean lands (Chapter 20), susceptible soils in the drier portions of the region have been subject to accelerated wind erosion. This forms part of the phenomenon of desertification. Deforestation, high stocking levels of domestic animals, cultivation, and miscellaneous recreational pressures, have all helped to create this problem in North Africa (Sghaier and Seiwert 1993), the Levant (Massri et al. 2002) and in the semi-arid lands of Spain (Lopez et al. 2001). However, the GLASOD (Global Assessment of Soil Degradation) survey of wind erosion severity (Middleton and Thomas 1997: 32–3) suggests that at present, with the exception of parts of North Africa, the Levant, and Sicily, wind erosion severity is generally low in the Mediterranean region. In addition, the Mediterranean lands (Figure 14.1) are in close proximity to the world’s greatest arid zone— the Sahara-Arabian belt—and so are subject to dust incursions from winds from that region: the ‘ghibli’ of Tripolitania, the ‘chili’ of Tunisia, the ‘khamsin’ of Egypt, and the ‘sirocco’ and ‘leveche’ of southern Europe. This has important geochemical implications (Kocak et al. 2004a, b). Knowledge of the dynamics of aeolian dust and sand transport comes from two main sources. The first of these is contemporary process monitoring, including data from dust traps, climatological stations, and remote sensing. The second is the long-term sedimentary record from such environments as caves, the sea-floor, lakes,

bogs, and loess deposits. There are, however, problems with gaps in the stratigraphic record, and uncertainties and limitations with respect to developing accurate geochronologies.

Dust Atmospheric dust comprised of mineral aerosol derived by deflation of desert surfaces, much of it from the Sahara (Middleton and Goudie 2001; Goudie and Middleton, 2006), is a feature of the Mediterranean basin, and it impacts upon the environment in a number of ways (Goudie and Middleton 2001). First, Saharan dust aerosols influence the nutrient dynamics and biogeochemical cycling of both oceanic and terrestrial ecosystems and have a major influence on soil characteristics (including terra rossa), oceanic productivity, air chemistry (Kubilay et al. 2003), and PM10 levels (Dordevic 2004). Second, atmospheric dust loadings may have considerable climatic significance, affecting rainfall amounts (Levin et al. 1996) and air temperatures through the absorption and scattering of solar radiation (Haywood et al. 2003). Third, there is evidence that dust loadings can change substantially in response to climatic changes both in the long term (e.g. during various phases of the Pleistocene) (Mahowald et al. 1999) and in the short term (e.g. in response to the North Atlantic Oscillation) (Moulin et al. 1997). Saharan dust is often deposited in precipitation over southern Europe and has been reported since ancient times (Bücher and Lucas 1984). Individual events can be of a high magnitude, such as the dust fall in early March 1991 that covered an area of at least 320,000 km2 , stretching from Sicily in the south to Fennoscandia in the north (Bücher and Dessens 1992; Franzen et al. 1995).

416

Andrew Goudie Main sand seas >15 dust storm days per year

Po Valley Ebr Ebro Basin

ranada Granad Granada Loes Loess

KO KONY NYA KONYA

TUNI TUNIS

Chotts Grand Erg Er Occidental

Matmata Loess Loes

Main trajectories trajectorie of spring dus dust transport EL ARISH ARIS

Er TRIPOLI Grand Erg TRIPOL Orienta Oriental

Erg er Raoui Raou

CAIRO CAIR

Nege Negev Sina Sinai

Qattara Depressio Depression

Er Erg Iguidi

Er Erg

Er Erg ssaouane Issaouane

Edeye Calanscio Edeyen Calansci Ubar bari Ubari

0

500 km

Fig. 14.1. A map of some aeolian phenomena and locations in the Mediterranean basin.

Dust from North African sources has been collected from the atmosphere over the Mediterranean (Chester et al. 1984; Tomadin et al. 1984; Blanco et al. 2003) and several authors have noted its significant geological role as an input to Mediterranean deep-sea sediments (Tomadin 1974; Löye-Pilot et al. 1986; Tomadin and Lenaz 1989; Box et al. 2008). Guerzoni et al. (1999: 147) have suggested that ‘Both the magnitude and the mineralogical composition of atmospheric dust inputs indicate that aeolian deposition is an important (50 per cent) or even dominant (>80 per cent) contribution to sediments in the offshore waters of the entire Mediterranean basin.’ Aerosols may include appreciable iron concentrations that may influence marine biogeochemistry (Guieu et al. 2002). It has also been shown that dust inputs to the marine environment can stimulate bacterial activity in the water column. Pulido-Villena et al. (2008) have argued that the connections between dust inputs and the ocean carbon cycle may be more significant than previously thought. Dust may also be a primary source, through accumulation and weathering, for the formation of terra rossa soils on limestones (Yaalon and Ganor 1979; Macleod 1980; Rapp 1984),

though this is a controversial issue (Chapter 6). It is possible, for example, that limestone dissolution contributes to the clay fraction, though grain size and mineralogical analyses of insoluble residues are suggestive of the importance of aeolian inputs from distant or more local sources (van Andel 1998). The Saharan source strength for dust transport to Europe was estimated at 80–120 × 106 tons per year by D’Almeida (1986) based on sunphotometer readings taken in the early 1980s. A major source area for transport to western Europe was identified in southernmost Algeria, between the Hoggar Massif and Adrar des Iforhas (Figure 14.1) Another source, where material is particularly rich in palygorskite (Molinaroli 1996), is in western Sahara/southern Morocco. These sources have been essentially confirmed by back trajectory analysis for dust deposited over north-eastern Spain. Avila et al. (1997) traced deposition events back to three main areas: western Sahara, Moroccan Atlas, and central Algeria. An area south-east of Benghazi in Libya is also an important source region (Koren et al. 2003). Transport to southern Europe is frequent. A year of monitoring on Corsica, for example, revealed twenty

Aeolian Processes

dust events (Bergametti et al. 1989) originating in three source areas: eastern Algeria/Tunisia/western Libya, Morocco/western Algeria, and ‘south of 30◦ N . Dust transport has also been monitored through analysis of satellite borne Total Ozone Mapping Spectrometer (TOMS) data (e.g. Kalivitis et al. 2007). The basis for this method is described by Herman et al. (1997) and Hsü et al. (1999). Data for the sample year of 1999 show that dust penetrated the troposphere over the Mediterranean on more than 60 per cent of days throughout the months of March to September, with Mediterranean dust outbreaks recorded on 100 per cent of days in June and August (Table 14.1a) (Middleton and Goudie 2001). In the western Mediterranean, a swathe of Saharan material penetrated as far as Sardinia on 43 per cent of days in June and 60 per cent of days in August (Table 14.1b). The most frequent source for dust reaching Sardinia was central-southern Tunisia and adjacent areas of eastern Algeria, an area of salt pans or chotts shown by Dubief (1953) as generating more than forty vents de sable. Morocco/western Algeria was an TABLE 14.1. Dust over the Mediterranean Month

No. of dust events∗

Total observations

%

(a) Mediterranean basin January February March April May June July August September October November December Year

6 6 22 24 30 30 24 30 19 9 4 3

28 27 31 28 31 30 30 30 30 30 27 31

21.43 22.22 70.97 85.71 96.77 100.00 80.00 100.00 63.33 30.00 14.81 9.67

207

353

61.79

(b) Sardinia January February March April May June July August September October November December

1 0 3 4 11 13 9 18 4 2 0 0

28 27 31 28 31 30 30 30 30 30 22 31

3.57 0.00 9.68 14.29 35.48 43.33 30.00 60.00 13.33 6.67 0.00 0.00

Year

65

353

19.40

∗

Days with an Aerosol Index (AI) > 1.9.

Source: 1999 TOMS data in Middleton and Goudie (2001).

417

occasional source of dust reaching Sardinia according to the 1999 TOMS data, as was north-eastern Libya, the latter a source not mentioned by Bergametti et al. (1989) in their Corsican work. Analysis of meteorological station data by Middleton (1986) showed a broad area across eastern Algeria and western Libya with >15 dust storm days a year on average and three stations in north-eastern Libya with a long-term (1956–77) mean of >10 dust storm days a year. According to TOMS, the least active months in 1999 for the whole Mediterranean were November and December, both showing less than 15 per cent of days with material over some part of the Mediterranean. No outbreaks reached Sardinia in November, December, or February, and occurred on just one day in January (Middleton and Goudie 2001). Since the source strength and transport of Saharan dust to Europe is heavily dependent upon climatic parameters, variations in dust fall frequencies could be an indicator of climatic change and this aspect has attracted the attention of several studies in recent years. However, year-to-year variability is high, as shown by eleven years of deposition records (1984–94) for Corsica where the annual input of Saharan dust varied between 4.0 and 26.2 g m−2 over this period (Löye-Pilot and Martin 1996). Nonetheless these authors noted that annual deposition rates over Corsica were seen to peak in the 1980s, and declined in the first half of the following decade. Deposition over Corsica shows a bimodal seasonality (spring and autumn maxima) similar to that recorded in north-eastern Spain where a decline in the frequency of red rain events in the 1990s has been noted after a peak of activity in the late 1980s (Avila et al. 1997; Avila and Peñuelas 1999). Several other studies have remarked upon the peak in Saharan dust falls over Europe in the late 1980s. Dessens and Van Dinh (1990) noted a marked increase in the frequency of Saharan dust outbreaks depositing at the Midi-Pyrenees Aerology Observatory in Lannemezan, France, over the period 1983–9, a rise confirmed from a much longer dataset (1949–94) from the Spanish Mediterranean coast south of Alicante (Sala et al. 1996). The long-term average there was approximately 2 dust-rain days per year, but from 1985 to 1994 the annual total averaged 6.5 dust-rain days, with 9.0 dust-rain days per year recorded for the period 1989–94. A significant increase in the quantities of Saharan dust falling over the French Alps since the early 1970s, with very high inputs occurring after 1980 (De Angelis and Gaudichet 1991), was detected from an ice core that yielded dust deposition data for a

418

Andrew Goudie

thirty-year period (1955–85). Contrary to this evidence, however, Conte et al. (1996) show a decline in the frequency of strong siroccos over the period 1951–90 at Trapani in Sicily, probably due to an increase in anticyclonic activity in the western and central parts of the Mediterranean basin which tends to counteract the occurrence of frontal disturbances that generate the strong, dust-laden southerly winds from the Sahara (Chapter 3).

Eastern Mediterranean Trajectories Transport from North Africa to the eastern Mediterranean occurs predominantly during spring and is commonly associated with the eastward passage of frontal low pressure systems, while dust from sources in the Middle East is more typically transported to the Mediterranean in the autumn (Dayan 1986; Kubilay et al. 2000). Analysis of 23 heavy dust falls in Israel over a 20-year period suggest that the North African type is by far the most common (Ganor et al. 1991), and North African dust is also the main source in Turkey (Mace et al. 2003). Long-range transport of dust to the eastern Mediterranean from the Arabian Desert also occurs in events that tend to last for one day, whereas the transport of Saharan dust to the central Mediterranean is characterized by events lasting 2–4 days (Dayan et al. 1991). There is some seasonal variation in the source areas of dust reaching Israel, with Chad being the spring source, Egypt and the Red Sea the source in July/August, and Libya in the autumn (Israelevich et al. 2003). Central Algeria is identified by Ganor et al. (1991) as the most frequent source area for Saharan dust reaching Israel, while Ganor and Foner (1996) distinguish between material commonly transported from sources in the Hoggar and Tibesti mountains in northern Chad, the latter also picking up material from the Western and Sinai Deserts. Some 25 × 106 tons of Saharan dust were estimated by Yaalon and Ganor (1979) to reach the eastern Mediterranean basin annually, most settling into the Mediterranean Sea. This figure has subsequently been revised upward, to 70 × 106 tons per year (Ganor and Mamane 1982) and more recently to 100 × 106 tons per year (Ganor and Foner 1996). The increase reflects the steady rise in frequency of Saharan dust episodes over Tel Aviv from ten per year in 1958 to nineteen per year in 1991 (Ganor 1994). The passage of Saharan dust into the Levant is of relatively frequent occurrence. The type of synoptic situation that leads to this is the passage of an advancing cold

front and the associated strong surface winds ahead of it penetrating south-eastwards from the Mediterranean Sea deep into the northern Sahara and Libyan Desert (Michaelides et al. 1999). This can be illustrated by the situation in mid-March 1998 (Goudie and Middleton 2002), when dust proved to be a significant natural hazard. A major dust event caused ports and airports to be closed, created breathing problems for inhabitants of Amman in Jordan, and led to fatal motoring accidents in Egypt and Jordan. Mean visibility in Amman was reduced to 4.2 km. A large, deep depression moved eastwards from North Africa, and then deepened further over the Middle East as it encountered cold polar air pushing across Turkey. The progress of this system can be traced by looking at the daily TOMS maps for the period from 14–19 March (Figure 14). The sequence starts with an area in eastern Algeria, southern Tunisia, and north-western Libya generating Aerosol Index (AI) values greater than 2.6. This index is a semi-quantitative measure of the amount of dust aerosol in the atmosphere, with higher index values generally indicating higher dust loadings. Clean air has a value of 0 and as a rule of thumb dusty air has values that exceed 2.0. The following day the dust pall has moved across Libya into Egypt and the eastern Mediterranean. On 16 March the main area with high AI values covers Cyprus, Israel, Syria, Jordan, and Sinai. Mean visibility at Amman airport was reduced to 3.2 km. On 17 March the area with high AI values has broken down into a group of small, deep clusters and by 19 March most of the area has values less than 1.0. Yet another example of the movement of a dust storm from the northern Sahara to the Middle East is provided by the events of April 2000 (Figure 14.3). On 18 April TOMS shows a large area of dust over southern Libya and far western Egypt. On the following day this reached eastern Egypt, Israel, and Lebanon, causing the closure of the port of Alexandria and the cessation of flights to Aswan. The mean visibility at Cairo airport was reduced to 4.6 km and that at Luxor to 3.7 km. In Israel mean visibility in Beersheva was reduced to 2.6 km. Limassol in Cyprus was also badly affected, as were flights in southern Turkey. The dust event than moved over Jordan, where the level of activity was less. Rates of dust deposition decline with distance from source (Table 14.2). Thus the annual values for western Europe (e.g. Central France and the Alps) are less Fig. 14.2. (opposite) The passage of dust systems from North Africa to the Middle East, mid March 1998, based on TOMS (from Goudie and Middleton 2002). The scale shows Toms Aerosol Index (AI) values.

14/3/98

15/3/98

16/3/98

17/3/98

18/3/98

19/3/98

1.0 1.4 1.8 2.2 2.6 3.0 3.4

420

Andrew Goudie 17/4/00

18/4/00

19/4/00

0.7

1.1

1.5

1.9

2.3

2.7

3.1

TOMS Aerosol Index (AI) High ground

Fig. 14.3. The TOMS sequence across North Africa to the Middle East for mid April, 2000 (from Goudie and Middleton 2002).

then 1 g m−2 . Further south, in north-east Spain, a value of 5.1 g m−2 is recorded, while over Sardinia, Corsica, Crete, and the south-east Mediterranean, most values are between 10 and 40 g m−2 . Included in the dust deposits may be such clay minerals as palygorskite (Molinaroli 1996; Tomadin et al. 1984) and kaolinite (Foucault and Mélières 2000; Avila et al. 1996). These may reflect different source areas, with the former tending to imply a more arid origin than the latter. Krom et al. (1999) used strontium isotopes as a tracer to investigate the relative contribution of Nile suspended sediments and Saharan dust inputs to the eastern

Mediterranean Sea, while Frumkin and Stein (2004) used strontium and uranium isotopes to identify Saharan dust in speleothems in Jerusalem. Grain size data for dust are provided in Table 14.3. These indicate that most of the dust in the Mediterranean is fine to medium silt.

Loess As already mentioned, dust deposition has contributed to the nature of both terrestrial and marine sediments. It has also been found in lake sediments (Narcisi 2000). The role of dust is especially important in the case

Aeolian Processes TABLE 14.2. Dust deposition amounts across the Mediterranean Source

Nihlen and Olsson (1995) Le-Bolloch et al. (1996) Wagenbach and Geis (1989) De Angelis and Gaudichet (1991) Avila et al. (1996) Bergametti et al. (1989) Löye-Pilot et al. (1986) Bücher and Lucas (1984) Pye (1992) Herut and Krom (1996) Herut and Krom (1996) Hernández and Hernández (1997)

Location

Aegean Sea Southern Sardinia Swiss Alps French Alps NE Spain Corsica Corsica Central France Crete Israel coast SE Mediterranean SE Spain

Annual deposition (g m−2 ) 11.2–36.5 6–13 0.4 0.2 5.1 12 12.5 1 10–100 72 36 23

of aeolian silts (loess), deposits of which are known from various parts of the Mediterranean basin (CoudéGaussen 1990). The most notable deposits are probably those from southern Tunisia (Figure 14.4) (Brunnacker 1973;

421

TABLE 14.3. Particle size characteristics of dust in various parts of the Mediterranean Source

Coudé-Gaussen (1991) Mattsson and Nihlen (1996) Sala et al. (1996) Pye (1992) Ozer et al. (1998) Bücher and Lucas (1994) Coudé-Gaussen (1991) Tomadin et al. (1984) Blanco et al. (2003)

Location

Modal, mean or median size in microns (Ïm)

Maghreb Crete Spain Crete Genoa, Italy SW France South of France Central Mediterranean Southern Italy

5–40 (median) 8–30 (modal) 4–30 (mean) 4–16 (median) 14.6 (median) 4–2.7 (median) 8–11 (median) 2–8 (modal) 1.7–2.4 (median)

Coudé-Gaussen et al. 1982), where the Matmata plateau loess reaches a thickness of 18 m at Téchine and contains up to five palaeosols typically rich in smectite and palygorskite. The loess probably derives from the Chott Djerid (a large ephemeral lake to the west) and from the Grand Erg Oriental (a large sand sea immediately to the

Fig. 14.4. The loess of Matmata, southern Tunisia, has been inhabited by cave dwellers (photo: Andrew Goudie).

422

Andrew Goudie

south-west). Coudé-Gaussen et al. (1983) suggest that two great phases of deposition occurred between 28,000 and 10,000 BP and from 6,000 to 4,000 BP, and CoudéGaussen et al. (1983) believed that maximum loess deposition occurred during humid conditions. This is a view that was disputed by Dearing et al. (1996) following their investigations of the mineral magnetic properties of these deposits. They believed that the period between 15,000 and 20,000 years BP was a time both of aridity and accelerated loess deposition. More recently, using luminescence dating, Dearing et al. (2001) have shown that some of the loess is older than this, with a sequence of loess and palaeosols from Téchine being deposited during the period 100,000–250,000 years BP. Loess has been known from Libya for many years (Schwegler 1944). Two loess deposits in Libya have also been recently studied. These have been classified as silty loess in the Tripoli region in the north-west of the country and clayey loess in the Ghat area in the southwest (Assallay 1996). Other significant loess deposits occur in Sinai (Rögner and Smykatz-Kloss 1991), but the most extensive deposits in the Middle East are those of the Negev (Yaalon and Dan 1974). Late Pleistocene loess, up to 10 m thick, occurs in the Granada basin of south-east Spain (Günster et al. 2001). Other loess is known from the central Apennines of Italy (Frezotti and Giraudi 1990), the Po Valley (Busacca and

Cremaschi 1998; Castiglioni 2001), Susak Island in the Dalmatian Archipelago (Cremaschi 1990), and in parts of Greece, including Crete (Brunnacker 1980). In most cases it is still not clear just to what degree the different loess deposits are derived from local or more distant processes, and to what extent some are derived from deserts and ephemeral lakes and some from formerly glaciated areas with large braided stream systems. Recent work in the Negev of southern Israel by Crouvi et al. (2008) has argued that the coarse silts of hilltop loess sequences were derived from the sand seas that advanced into Sinai and the Negev during the Late Pleistocene when the Mediterranean shelf was exposed.

Dunes and Other Aeolian Forms Because of the dryness of parts of the Mediterranean region, sand dunes have formed in various parts of the region, including its coastlines and arid interiors. Aeolian sand has also accumulated in various caves, providing important information about palaeoenvironments associated with archaeological sites (see e.g. Farrand 1979), especially in coastal settings. Some of the sand deposits may be the product of past more arid conditions, as is the case with the Plio-Pleistocene sands

Fig. 14.5. Barchan dunes in the Libyan Desert, Kharga depression, Egypt (photo: Andrew Goudie).

Aeolian Processes

of the Valdarno Basin of the northern Apennines in Italy (Ghinassi et al. 2004). Likewise, in Mallorca, dune sand appears to have been transported from exposed carbonate shelves far inland by westerly winds under cold, dry, windy phases of the Pleistocene (Nielsen et al. 2004). One extensive dune area occurs on the Mediterranean coast of Sinai, in the vicinity of El Arish. The so-called Sinai–Negev erg, which covers about 10,000 km2 , is formed of sand derived from the Nile, which has been reworked by the Mediterranean and blown inland from its beaches (Tsoar and Goodfriend 1994), and from the Mediterranean shelf at times of lowered sea level (Crouvi et al. 2008). Much of this erg, which contains both transverse and linear forms, developed between 19,000 and 10,000 years BP (Goring-Morris and Goldberg 1990), but there have been various phases of remobilization in the Holocene that could have resulted from anthropogenic pressures. Indeed, many of the coastal parabolic dunes of the Israeli and Palestine coast have formed

423

in the last 1,000 years, probably because of vegetation removal associated with agricultural and pastoral activities (Tsoar and Blumberg 2002). Only 17 per cent of the Israeli coastal dunes are still of good or reasonable ecological value (Kutiel 2001), and recreational impacts can be severe (Kutiel et al. 1999). Elsewhere in the Near East, some Late Pleistocene and Holocene dunes have been studied in the Konya plain of Turkey (Kuzucuoglu et al. 1998). Inland dunes are also known from the Po valley in northern Italy (Castiglioni 2001). Coastal dune systems are extensively developed around the Mediterranean, not least in the Balearic Islands, where vegetated parabolic forms are widespread (Servera and Rodriguez 1996), together with cliff-front ramps (Clemmensen et al. 2001). However, large tracts of them are being modified or destroyed by tourist pressures (Schmitt 1994; Chapter 13). In mainland Spain parabolic dunes are also common in Catalonia (Gos and Serra 1993) and Valencia (Sanjaume

Fig. 14.6. Gypsum crust soils with polygonal structures in southern Tunisia. Deflation of gypsum from dry lake basins and its deposition downwind contributes to the formation of these widespread surface material types (photo: Andrew Goudie).

424

Andrew Goudie

and Pardo 1991, 1992), but are also being subjected to intense human pressures. Perhaps the most celebrated dunes of Spain are, however, on its Atlantic coast in the Doñana National Park (Munoz-Reinoso 2001), but here too human activities (including over-exploitation of groundwater) pose a threat. A general review of sanddune management problems and techniques in Spain is provided by Gómez-Pina et al. (2002). France also has many dune systems along the coast of the Gulf of Lions (Corre 1991). Urban sprawl is a major threat to coastal dunes along the coast of Egypt, especially between Alexandria and El Alamein (Batanouny 1999). One of the most interesting features of many of the coastal dunes on the shorelines of the Mediterranean is the formation of lithified carbonate-rich dunes or aeolianites. These occur in mainland Spain, the Balearic Islands, Corsica, and Sardinia, Greece (Gittenberger and Goodfriend 1993), Cyprus and Crete (Kelletat 1991), North Africa (El-Asmar 1994; Plaziat and Mahmoudi 1990), and along the Levant coast. A bibliography is provided by Brooke (2001). In Israel this material is often called kurkar and contains multiple

palaeosols (hamra) indicating alternating stability and dune activity (Verrecchia 1989; Tsoar 2000; Frechen et al. 2002) during the Late Pleistocene (Sivan and Porat 2004). On the south side of the Mediterranean extensive dunes occur in Morocco, Algeria, Tunisia, Libya, and Egypt—the northern Saharan dunes. They form great sand seas, which from west to east are the Grand Erg Occidental, the Grand Erg Oriental, the Calanscio Erg, the Great Sand Sea, and the Abu Moharik belt (Goudie 2002) (Figures 14.1 and 14.5). These northern Saharan sand seas have complex patterns of dunes as a consequence of multidirectional wind regimes involving the interaction of both the northeasterly trades and the mid-latitude circulation (Chapter 3). It is, of course, difficult to say where the Mediterranean region stops and the Sahara starts, but undoubtedly large tracts of the northern Sahara have important impacts on the Mediterranean lands. In Tunisia and Algeria the presence of a series of closed depressions—the Chotts—which may themselves in part be the product of aeolian deflation from

Fig. 14.7. The great lunette dune on the lee side of the Sebkha el Kelbia, central Tunisia. The salt lake from which it is derived is visible in the background (photo: Andrew Goudie).

Aeolian Processes

structurally controlled basins (Sweezey 1996, 1997)— has promoted the development of gypsum dunes and has contributed to the formation of some gypsum crusts (Figure 14.6) (Watson 1985) and of some enormous lunettes on their lee sides (Coque 1979; Perthuisot and Jauzein 1975; Goudie and Wells 1995; Benazzouz 1986). In the case of the Sebkha el Kelbia in Tunisia the lunette is 146 m high (Figure 14.7). In Egypt, just to the south of the Mediterranean coast, is the Qattara Depression, the base of which lies at −133 m below present sea level. The origin of this 19,500 km2 depression has been the subject of considerable controversy. Some workers have seen it as essentially deflational, others as being tectonic, others as being related to old river courses, and others as being the result of a combination of deflation and salt weathering (Goudie 2002; Aref et al. 2002; Ball 1927). Another area of closed drainage depressions in the Mediterranean region is the Ebro basin of Spain (mean annual rainfall 350 mm). Here there are a large number of closed depressions, called ‘saladas’. Their origin partly results from solution of Tertiary evaporite strata, but they have been deepened and extended by deflation (Sánchez et al. 1998). Yardangs (wind fluted rock) and nebkha dunes (which accumulate around vegetation) have developed in the same area (Gutiérrez-Elorza et al. 2002). Interestingly, clay dunes, derived from deflated saline playas, occur in the Guadiana valley of Spain (Rendell et al. 1994). Thus we have the whole spectrum of aeolian particle sizes in dune systems in the Mediterranean region. Finally, wind action plays a major role in the dispersal of volcanic tephra across the region (Chapter 15) and its subsequent deposition on land and in the sea (Narcisi and Vezzoli 1999) (Chapter 2). It is also responsible for the transport of large quantities of anthropogenic atmospheric pollutants into the area from Europe, Asia, and beyond (Lelieveld et al. 2002) (Chapter 22).

Conclusion Saharan dust has a major impact on the climate, biogeochemical cycling, atmospheric quality, and soils of the Mediterranean countries. It also constitutes a significant natural hazard. In the Pleistocene, the Sahara and other sources led to the development of scattered but locally extensive loess deposits, most notably in Tunisia and the Negev. Also important has been the development of coastal dunes, many of them now under severe anthropogenic pressures, and of larger continental dune

425

systems, especially in North Africa and the Negev/Sinai region. Wind erosion has created erosional features such as deflation basins and yardangs in North Africa and semi-arid Spain. Wind erosion and dune degradation are being exacerbated at the present time by burgeoning tourism and recreation, by deforestation, cultivation, overgrazing, and the lowering of water tables, but overall it is not as potent a threat as that posed by water erosion and slope destabilization.

References Aref, M. A. M., El-Khoriby, E., and Hamdan, M. A. (2002), The role of salt weathering in the origin of the Qattara Depression western Desert, Egypt. Geomorphology 45: 181–95. Assallay, A. M., Rogers, C. D. F., and Smalley, I. J. (1996), Engineering properties of loess in Libya. Journal of Arid Environments 32: 373–86. Avila, A. and Peñuelas, J. (1999), Increasing frequency of Saharan rains over northeastern Spain and its ecological consequences. The Science of the Total Environment 228: 153–6. Queralt, I., Gallart, F., and Martin-Vide, J. (1996), African dust over northeastern Spain: mineralogy and source regions, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 201–15. Queralt-Mitjans, I., and Alarcón, M. (1997), Mineralogical composition of African dust delivered by red rains over northeastern Spain. Journal of Geophysical Research 102 (D18): 21,977–96. Ball, J. (1927), Problems of the Libyan Desert. Geographical Journal 70: 21–8, 105–28, 209–24. Batanouny, K. H. (1999), The Mediterranean coastal dunes in Egypt: an endangered landscape. Estuarine Coastal and Shelf Science 49 (Suppl. A): 3–9. Benazzouz, M. (1986), Recherches géomorphologiques dans les hautes plaines de l’est Algérien: la sebkhat Tarf (Algérie). Ph.D. Thesis. University of Paris I, Sorbonne. Bergametti, G., Gomes, L., Remoudaki, E., Desbois, M., Martin, D., and Buat Ménard, P. (1989), Present transport and deposition patterns of African dusts to the north-western Mediterranean, in M. Leinen and M. Sarnthein (eds.), Palaeoclimatology and Palaeometeorology: Modern and Past Patterns of Global Atmospheric Transport, NATO ASI Series C. 282: 227–52. Blanco, A., De Tomasi, F., Filippo, E., Manno, D., Perrone, M. R., Serra, A., Tafuro, A. M., and Tepore, A. (2003), Characterization of African dust over southern Italy. Atmospheric Chemistry and Physics 3: 2147–59. Box, M. R., Krom, M. D., Cliff, R., Almogi-Labin, A., Bar-Matthews, M., Ayalon, A., Schillman, B., and Paterne, M. (2008), Changes in the flux of Saharan dust to the East Mediterranean Sea since the last glacial maximum as observed through Sr-isotope geochemistry. Mineralogical Magazine 72: 307–11. Brooke, B. (2001), The distribution of carbonate eolianite. EarthScience Reviews 55: 135–64. Brunnacker, K. (1973), Einiges über Löss-Vorkommen in Tunesien. Eiszeitalter und Gegenwart 23/4: 1–11.

426

Andrew Goudie

Brunnacker, K. (1980), Young Pleistocene loess as an indicator for the climate in the Mediterranean area. Palaeoecology of Africa 12: 99–113. Bücher, A. and Dessens, J. (1992), Poussières sahariennes sur la France et l’Angleterre, 6–9 mars 1991. Journal of Meteorology 17: 226–33. and Lucas, G. (1984), Sédimentation éolienne intercontinentale, poussières sahariennes et géologie. Bulletin Centre Researche et Exploration-Production Elf-Aquitaine 8: 151–65. Busacca A. and Cremaschi M. (1998), The role of time versus climate in the formation of the deep soils of the Apennine fringe of the Po Valley, Italy. Quaternary International 51/2: 95–107. Castiglioni, G. B. (2001), Forme e depositi di origine eolica. Supplementi di geografia fisica e dinamica quaternaria 4: 119–22. Chester, I., Sharples, E. J., Sanders, G. S., and Saydam, A. C. (1984), Saharan dust incursion over the Tyrrhenian Sea. Atmospheric Environment 18: 929–35. Clemmensen, L. B., Lisborg, T., Fornos, J. L., and Bromley, R. G. (2001), Cliff-front aeolian and colluvial deposits, Mallorca, western Mediterranean: a record of climatic and environmental change during the last glacial period. Bulletin of the Geological Society of Denmark 48: 217–32. Conte, M., Colacine, M., and Piervitali, E. (1996), Atlantic disturbances deeply penetrating the African continent: effects over Saharan regions and the Mediterranean Basin, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 93–102. Coque, R. (1979), Sur la place du vent dans l’érosion en milieu aride. L’example des lunettes (bourrelets éolian) de la Tunisie. Mediterranée 15–21. Corre, J.-J. (1991), The sand dunes and their vegetation along the Mediterranean coast of France: their likely response to climatic change. Landscape Ecology 6: 65–75. Coudé-Gaussen, G. (1990), The loess and loess-like deposits along the sides of the western Mediterranean sea: genetic and palaeoclimatic significance. Quaternary International 5: 1–8. (1991), Les poussières Sahariennes. John Libbey Eurotext, Montrouge. Mosser, C., Rognon, P., and Torenq, J. (1982), Une accumulation de loess Pleistocene supérieur dans le sud-Tunisien: coupe de Téchine. Bulletin de Société Géologique de France 24: 283–92. Olive, L., and Rognon, P. (1983), Datation de depots loessiques et variations climatiques à la bordure nord du Sahara algéro-tunisien. Revue de Géomorphologie Dynamique et de Géographie Physique 24: 61–73. Cremaschi, M. (1990), Stratigraphy and palaeoenvironnmental significance of the loess deposits on Susak Island (Dalmatian Archipelago). Quaternary International 5: 97–105. Crouvi, O., Amit, R., Enzel, Y., Porat, N., and Sandler, A. (2008), Sand dunes as a major proximal dust source for Late Pleistocene loess in the Negev Desert, Israel. Quaternary Research 70: 275–82. D’Almeida, G. A. (1986), A model for Saharan dust transport. Journal of Climate and Applied Meteorology 25: 903–16. Dayan, U. (1986), Climatology of back trajectories from Israel based on synoptic analysis. Journal of Climate and Applied Meteorology 25: 591–5. Heffter, J., Miller, J., and Gutman, G. (1991), Dust intrusion into the Mediterranean basin. Journal of Applied Meteorology 30: 1185–99.

De Angelis, M. and Gaudichet, A. (1991), Saharan dust deposition over Mont Blanc (French Alps) during the last 30 years. Tellus 43B: 61–75. Dearing, J. A., Livingstone, I. P., and Zhou, L. P. (1996), A late Quaternary magnetic record of Tunisian loess and its climatic significance. Geophysical Research Letters 23: 189–92. Bateman, M. D., and White, K. (2001), Palaeoclimate record from OIS 8.0–5.4 recorded in loess in palaeosol sequences on the Matmata Plateau, southern Tunisia, based on mineral magnetism and new luminescence dating. Quaternary International 76/7: 43–56. Dessens, J. and Van Dinh, P. (1990), Frequent Saharan dust outbreaks north of the Pyrenees: a sign of climatic change? Weather 45: 327–33. Dordevic, D., Vukmirovic, Z., Tosic, I., and Unkasevic, M. (2004), Contribution of dust transport and resuspension to particulate matter levels in the Mediterranean atmosphere. Atmospheric Environment 38: 3637–45. Dubief, J. (1953), Les vents de sable dans le Sahara Français. Colloques Internationaux de CNRS 35: 45–70. El-Asmar, H. M. (1994), Aeolianite sedimentation along the northwestern coast of Egypt: evidence for middle to late Quaternary aridity. Quaternary Science Reviews 13: 699–708. Farrand, W. R. (1979), ‘Chronology and palaeoenvironment of Levantine prehistoric sites as seen from sediment studies’. Journal of Archaeological Science 6: 369–92. Fornos, J. J., Bromley, R. G., Clemmensen, L. B., and RodriguezPerea, A. (2002), ‘Tracks and trackways of Myotragus balearicus bate (Artiodactyla, Caprinae) in Pleistocene aeolianites from Mallorca (Balearic Islands, western Mediterranean)’. Palaeogeography, Palaeoclimatology, Palaoecology 180: 277–313. Foucault, A. and Mélières, F. (2000), Palaeoclimatic cyclicity in central Mediterranean Pliocene sediments: the mineralogical signal. Palaeogeography, Palaeoclimatology, Palaeoecology 158: 311–23. Franzen, L. G., Hjelmroos, M., Kalberg, P., Rapp, A., Mattsson, J. O., and Brorström-Lundén, E. (1995), The Saharan dust episode of south and central Europe, and northern Scandinavia, March 1991. Weather 50: 313–18. Frechen, M., Neber, A., Dermann, B., Tsatskin, A., Boenigk, W., and Ronen, A. (2002), Chronostratigraphy of aeolianites from the Sharon Coastal Plain of Israel. Quaternary International 89: 31–44. Frezotti, M. and Giraudi, C. (1990), Late glacial and Holocene aeolian deposits and geomorphological features near Roccaraso (Abruzzo-central Italy). Quaternary International 5: 89–95. Frumkin, A. and Stein, M. (2004), The Sahara-East Mediterranean dust and climate connection revealed by strontium and uranium isotopes in a Jerusalem speleothem. Earth and Planetary Science Letters 217: 451–64. Ganor, E. (1994), The frequency of Saharan dust episodes over Tel Aviv, Israel. Atmospheric Environment 28: 2867–71. and Foner, H. A. (1996), The mineralogical and chemical properties and the behaviour of aeolian Saharan dust over Israel, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 163–72. and Mamane, M. (1982), Transport of Saharan dust across the eastern Mediterranean. Atmospheric Environment 16: 581–7.

Aeolian Processes Ganor, E., Foner, H. A., Brenner, S., Neeman, E., and Lavi, N. (1991), The chemical composition of aerosols settling in Israel following dust storms. Atmospheric Environment 25: 2665–70. Ghinassi, M., Magi, M., Sagri, M., and Singer, B. S. (2004), Arid climate 2.5 Ma in the Plio-Pleistocene Valdarno Basin (Northern Apennines, Italy). Palaeogeography, Palaeoclimatology, Palaeoecology 207: 37–57. Gittenberger, E. and Goodfriend, G. A. (1993), Landsnails from the Last Glacial Maximum on Andikithira, southern Greece and their palaeoclimatic implications. Journal of Quaternary Science 8: 109–16. Gómez-Pina, G., Muñoz-Pérez, J. J., Ramirez, J. L., and Ley, C. (2002), Sand dune management problems and techniques, Spain. Journal of Coastal Research Special Issue 36: 325–32. Goring-Morris, A. N. and Goldberg, P. (1990), Late Quaternary dune incursions in the southern Levant: archaeology, chronology and palaeoenvironments. Quaternary International 5: 115– 37. Gos, L. and Serra, J. (1993), A complex dune system in Baix Emporda (Catalonia, Spain). Geological Society of London Special Publication 72: 191–9. Goudie, A. S. (2002), Great Warm Deserts of the World: Landscapes and Evolution. Oxford, Oxford University Press. and Middleton, N. J. (2001), Saharan dust storms; nature and consequences. Earth-Science Reviews 56: 179–204. and Wells, G. L. (1995), The nature, distribution and formation of pans in arid zones. Earth-Science Reviews 38: 1–69. (2002), Dust storms of the Middle East. Bulletin de la Classe des Sciences, Académie Royale de Belgique, ser. 6, 13: 379–99. (2006), Desert Dust in the Global System. Springer, Berlin. Guerzoni, S., Chester, R., Dulac, F., Herut, B., Löye-Pilot, M. D., Measures, C., Mignon, C., Molinari, E., Moulin, C., Rossini, P., Saydam, C., Soudine, A., and Ziveri, P. (1999), The role of atmospheric deposition in the biogeochemistry of the Mediterranean Sea. Progress in Oceanography 44: 147–90. Guieu, C., Bozec, Y., Blain, S., Ridame, C., Sarthou G., and Leblond, N. (2002), Impact of high Saharan dust inputs on dissolved iron concentrations in the Mediterranean Sea. Geophysical Research Letters 29/19: 1911, doi: 10.1029/2001GLO14454. Günster, N., Eck, P., Skowronek, A., and Zöller, L. (2001), Late Pleistocene loess and their paleosols in the Granada Basin, southern Spain. Quaternary International 76/7: 241–5. Gutiérrez-Elorza, M., Desir, G., and Gutiérrez-Santolalla, F. (2002), Yardangs in the semiarid central sector of the Ebro Depressions (NE Spain). Geomorphology 44: 155–70. Haywood, J., Francis, P., Osborne, S., Glew, M., Loeb, N., Highwood, E., Tanre, D., Myhre, G., Formenti, P., and Hirst, E. (2003), Radiative properties and direct radiative effect of Saharan dust measured by the C-130 aircraft during SHADE: 1. Solar spectrum. Journal of Geophysical Research 108 (D18): 8577, doi: 10.1029/2002JD002687. Herman, J. R., Bhartia, P. K., Torres, O., Hsu, C., Seftor, C., and Celarier, E. (1997), Global distribution of UV-absorbing aerosols from Nimbus 7 TOMS data. Journal of Geophysical Research 102 (D14): 16911–22. Hernández, J. L. D., and Hernández, J. M. M. (1997), Tasas de deposicion de polvo atmosferico en un area semiarida del entorno mediterraneao occidental. Estudios Geologicos 53: 211–20.

427

Herut, B. and Krom, M. (1996), Atmospheric input of nutrients and dust to the SE Mediterranean, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 349–58. Hsü, N. C., Herman, J. R., Torres, O., Holben, B., Tanre, D., Eck, T. F., Smirnov, A., Chatenet, B., and Lavenu, F. (1999), Comparison of the TOMS aerosol index with sun photometer aerosol optical thickness: results and applications. Journal of Geophysical Research 104: 6269–79. Israelevich, P. L., Ganor, E., Levin, Z., and Joseph, P. J. (2003), Annual variations of physical properties of desert dust over Israel. Journal of Geophysical Research 108 (D13): article 4381. Kalivitis, N., Gerasopoulos, E., Vrekoussis, M., Kouvarakis, G., Kubilay, N., Hatzianastassiou, N., Vardavas, I., and Mihalopoulos, N. (2007), Dust transport over the eastern Mediterranean derived from Total Ozone Mapping Spectrometer, Aerosol Robotic Network, and surface measurements. Journal of Geophysical Research 112: D03202, doi: 10.1029/2006JD007510. Kelletat, D. (1991), The 1550 BP tectonic event in the eastern Mediterranean as a basis for assessing the intensity of shore processes. Zeitschrift für Geomorphologie Suppl. 81: 181–94. Kocak, M., Nimmo, M., Kublay, N., and Herut, B. (2004a ), Spatiotemporal aerosol trace metal concentrations and sources in the Levantine basin of the Eastern Mediterranean. Atmospheric Environment 38: 2133–44. Kubilay N., and Mihalopoulos, N. (2004b), Ionic composition of lower tropospheric aerosols at a Northeastern Mediterranean site implications regarding sources and long-range transport. Atmospheric Environment 38: 2067–77. Koren, I., Joseph, J. H., and Israelevich, P. (2003), ‘Detection of dust plumes and their sources in northeastern Libya’. Canadian Journal of Remote Sensing 29: 792–6. Krom, M. D., Cliff R. A., Eijsink L. M., Herut, S., and Chester, R. (1999), The characterisation of Saharan dusts and Nile particulate matter in surface sediments from the Levantine basin using Sr isotopes. Marine Geology 155: 319–30. Kubilay, N., Nickovic, S., Moulin, C., and Dulac, F. (2000), An illustration of the transport and deposition of mineral dust onto the eastern Mediterranean. Atmospheric Environment 34: 1293–303. Cokacar, T., and Oguz, T. (2003), Optical properties of mineral dust outbreaks over the northeastern Mediterranean. Journal of Geophysical Research 108: D21, AAC4, 1–9. Kutiel, P. (2001), Conservation and management of the Mediterranean coastal sand dunes in Israel. Journal of Coastal Conservation 7: 183–92. Zhevelev, H., and Harrison, R. (1999), The effect of recreational impacts on soil and vegetation of stabilised coastal dunes in the Sharon Park, Israel. Ocean and Coastal Management 42: 1041–60. Kuzucuoglu, C., Parish, R., and Karabiyikoglu, M. (1998), The dune systems of the Konya Plain (Turkey): Their relation to environmental changes in Central Anatolia during the Late Pleistocene and Holocene. Geomorphology 23: 257–21. Le-Bolloch, O., Guerzoni, S., and Molinaroli, E. (1996), Atmosphere-ocean mass fluxes at two coastal sites in Sardinia 39–41 degrees N, 8–10 degree E’, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 217–22. Lelieveld, J. and 30 others (2002), Global air pollution crossroads over the Mediterranean. Science 298: 794–9.

428

Andrew Goudie

Levin, Z., Ganor, E., and Gladstein, V. (1996), The effects of desert particles with sulphate on rain formation in the Eastern Mediterranean. Journal of Applied Meteorology 35: 1511–23. Lopez, M. V., Gracia, R., and Arrue, J. L. (2001), An evaluation of wind erosion hazard in fallow lands of semi-arid Aragon (NE Spain). Journal of Soil and Water Conservation 56: 212–19. Löye-Pilot, M. D. and Martin, J. M. (1996), Saharan dust input to the western Mediterranean: an eleven year record in Corsica, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 191–9. and Morelli, J. (1986), Influence of Saharan dust on the rainfall acidity and atmospheric input to the Mediterranean. Nature 321: 427–8. Mace, K. A., Kubilay, N., and Duce, R. A. (2003), ‘Organic nitrogen in rain and aerosol in the eastern Mediterranean atmosphere: an association with atmospheric dust’. Journal of Geophysical Research 108: article 4320. Macleod, D. A. (1980), The origin of the red Mediterranean soils in Epirus, Greece. Journal of Soil Science 31: 125–36. Mahowald, N., Kohfeld, K., Hansson, M., Balkanski, Y., Harrison, S. P., Prentice, I. C., Schulz, M., and Rodhe, H. (1999), Dust sources and deposition during the Last Glacial Maximum and current climate: a comparison of model results with paleodata from ice cores and marine sediments. Journal of Geophysical Research 104 (D13): 15895–916. Massri, Z., Zöbisch, M., Bruggeman, A., Hayek, P., and Kardous, M. (2002), Wind erosion in marginal Mediterranean dryland areas, Khanasser Valley—a case study, in J. A. Lee and T. M. Zobeck (eds.), Proceedings of ICAR5/GCTE-SEN Joint Conference, Lubbock, 174–9. Mattsson, J. O. and Nihlen, T. (1996), The transport of Saharan dust to southern Europe: a scenario. Journal of Arid Environments 32: 101–10. Michaelides, S., Evripidou, G., and Kallos, G. (1999), Monitoring and predicting Saharan Desert dust events in the eastern Mediterranean. Weather 54: 359–65. Middleton, N. J. (1986), Effect of drought on dust production in the Sahel. Nature 316: 431–4. and Goudie, A. S. (2001), Saharan dust: sources and trajectories. Transactions of the Institute of British Geographers NS 26: 165–81. and Thomas, D. S. G. (1997), World Atlas of Desertification, 2nd edn. Arnold, London. Molinaroli, E. (1996), Mineralogical characterisation of Saharan dust with a view to its final destination in Mediterranean sediments, in S. Guerzoni and R. Chester (eds.), The Impact of Desert Dust across the Mediterranean. Kluwer, Dordrecht, 153–62. Moulin, C., Lamber, C. E., Dulac, F., and Dayan, U. (1997), Control of atmospheric export of dust by the North Atlantic Oscillation. Nature 387: 691–4. Munoz-Reinoso, J. C. (2001), Vegetation changes and groundwater abstraction in SW Donana, Spain. Journal of Hydrology 242: 197–209. Narcisi, B. (2000), Late Quaternary eolian deposition in Central Italy. Quaternary Research, 54: 246–52. and Vezzoli, L. (1999), Quaternary stratigraphy of distal tephra layers in the Mediterranean—an overview. Global and Planetary Change 21: 31–50. Nielsen, K. A., Clemmensen, L. B., and Fornós, J. J. (2004), Middle Pleistocene magnetostratigraphy and susceptibility stratigraphy: data from a carbonate aeolian system, Mallorca,

Western Mediterranean. Quaternary Science Reviews 23: 1733–56. Nihlen, T., and Olsson, S. (1995), Influence of eolian dust on soil formation in the Aegean area. Zeitschrift für Geomorphologie 39: 341–61. Ozer, P., Erpicum, M., Cortemiglia, G. C., and Luchetti, G. (1998), A dustfall event in November 1996 in Genoa, Italy. Weather 53: 140–5. Perthuisot, J.-P. and Jauzein, A. (1975), ‘Sebkhas et dunes d’argile: l’enclave endoreique de Pont du Fahs, Tunisie. Revue de Géographie Physique et de Géologie Dynamique 17: 295–306. Plaziat, J.-C. and Mahmoudi, M. (1990), The role of vegetation in Pleistocene eolianite sedimentation: an example from eastern Tunisia. Journal of African Earth Sciences 10: 445–51. Pulido-Villena, E., Wagener, T., and Guieu, C. (2008), Bacterial response to dust pulses in the western Mediterranean: implications for carbon cycling in the oligotrophic ocean. Global Biogeochemical Cycles 22: GB1020, doi: 10.1029/ 2007GB003091. Pye, K. (1992), Aeolian dust transport and deposition over Crete and adjacent parts of the Mediterranean Sea. Earth Surface Processes and Landforms 17: 271–88. Rapp, A. (1984), Are terra rossa soils in Europe eolian deposits from Africa? Geologiska Föreningens i Stockholm Förhanlingar105: 161–8. Rendell, H. M., Calderon, T., Perez-Gonzalez, A., Gallardo, J., Millan, A., and Townsend, P. D. (1994), Thermoluminescence and optically stimulated luminescence dating of Spanish dunes. Quaternary Science Reviews 13: 429–32. Rögner, K. and Smykatz-Kloss, W. (1991), The deposition of eolian sediments in lacustrine and fluvial environments of Central Sinai (Egypt). Catena Suppl. 20: 75–91. Sala, J. Q., Cantos, J. O., and Chiva, E. M. (1996), Red dust within the Spanish Mediterranean area. Climatic Change 32: 215–28. Sánchez, J. A., Pérez, A., Coloma, P., and Martinez-Gil, J. (1998), Combined effects of groundwater and aeolian processes in the formation of the northernmost closed saline depressions of Europe: north-east Spain. Hydrological Processes 12: 813–20. Sanjaume, E. and Pardo, J. (1991), Dune regeneration of a previously destroyed dune field, Devesa del Saler, Valencia, Spain. Zeitschrift für Geomorphologie Suppl. 81: 125–34. (1992), The dunes of the Valencian coast (Spain): past and present. Proceedings 3rd European Dune Congress Galway, Balkema, Rotterdam, 475–86. Schmitt, T. (1994), Stress on and change in the sandy beach and dune ecosystems of Mallorca as a result of tourism. Geookodynamik 15: 165–85. Schwegler, E. (1944), Bemerkungen zum Vorkommen von Löss im Libyschen und tunesischen Gebiet. Neues Jahrbuch für Mineralogie Monatsheft B: 10–17. Servera, N. J. and Rodriguez, P. A. (1996), Parabolic morphology of the coastal dune systems of the Balearic Islands. Cadernos Laboratorio Xeoloxico de Laxe 21: 645–58. Sghaier, M. and Seiwert, W. D. (1993), Winds of change and the threat of desertification: case study from the Tunisian Sahara. GeoJournal 31: 95–9. Sivan, D. and Porat, N. (2004), Evidence from luminescence for Late Pleistocene formation of calcareous aeolianite (kurkar) and paleosol (hamra) in the Carmel Coast, Israel. Palaeogeography, Palaeoclimatology, Palaeoecology 211: 95–106.

Aeolian Processes Sweezey, C. S. (1996), Structural controls on Quaternary depocentres with the Chotts trough region of southern Tunisia. Journal of African Earth Sciences 22: 335–47. (1997), Climatic and tectonic controls on Quaternary eolian sedimentary sequences of the Chott Rharsa Basin, Southern Tunisia, PhD. Thesis. University of Texas at Austin. Tomadin, L. (1974), Les minéraux argileux dans le sediments actuels de la Mer Tyrréhnienne. Bulletin du Groupe Français des Argiles 26: 219–28. and Lenaz, R. (1989), Eolian dust over the Mediterranean and their contribution to the present sedimentation, in M. Leinen and M. Sarnthein (eds.), Palaeoclimatology and Palaeometeorology: Modern and Past Patterns of Global Atmospheric Transport. NATO ASI Series C. Vol. 282: 267–82. Landuzzi, V., Mazzucotelli, A., and Vannucci, R. (1984), Wind-blown dusts over the central Mediterranean. Oceanologica Acta 7: 13–23. Tsoar, H. (2000), Geomorphology and palaeogeography of sand dunes that have formed the kurkar ridges in the coastal plain of Israel. Israel Journal of Earth Sciences 49: 189–96. and Blumberg, D. G. (2002), Formation of parabolic dunes from barchan and transverse dunes along Israel’s Mediterranean coast. Earth Surface Processes and Landforms 27: 1147–61.

429

and Goodfriend, G. A. (1994), Chronology and palaeoenvironmental interpretation of Holocene aeolian sands at the inland edge of the Sinai-Negev erg. The Holocene 4: 244–50. van Andel, T. H. (1998), Paleosols, red sediments and the Old Stone Age in Greece. Geoarchaeology: An International Journal 13: 361–90. Verrecchia, E. (1989), Aeolianites (Kurkar) of the Israeli Mediterranean coast: digenetic evolution of Quaternary littoral sediments. Bulletin Centre de Géomorphologie du CNRS, Caen 36: 61–4. Wagenbach, D. and Geis, K. (1989), The mineral dust record in a high alpine glacier (Colle Gnifett, Swiss Alps), in M. Leinen and M. Sarnthein (eds.), Paleoclimatology and Paleometeorology: Modern and Past Patterns of Global Atmospheric Transport. Kluwer, Dordrecht, 543–64. Watson, A. (1985), Structure, chemistry and origin of gypsum crusts in southern Tunisia and the Central Namib Desert, Sedimentology 32: 855–75. Yaalon, D. H. and Dan, J. (1974), The influence of dust on soils during the Quaternary. Soil Science 116: 146–55. and Ganor, E. (1979), East Mediterranean trajectories of dust-carrying storms from the Sahara and Sinai, in C. Morales (ed.), Saharan Dust. Wiley, Chichester, 187–93.

This chapter should be cited as follows Goudie, A. S. (2009), Aeolian processes and landforms, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 415–429.

This page intentionally left blank

III

Hazards

This page intentionally left blank

Editorial Introduction Jamie Woodward

Catastrophic earthquakes, explosive volcanic eruptions, and devastating storms and floods are intimately bound up within the history and mythology of the Mediterranean world. It is a key region for the study of natural hazards because it offers unrivalled access to long records of hazard occurrence and impact through documentary, archaeological and geological archives. Early texts and archaeological data have provided unique insights into the nature and impact of past eruptions, earthquakes, tsunamis, and other hazards. Notable events were carefully documented in Antiquity and the archaeological record provides insights into the impact of catastrophic events on past human societies. The eruption of Vesuvius in AD 79, for example, was famously documented by Pliny the Younger, and the excavations at Pompeii have provided extraordinarily rich insights into the dynamics and impacts of tephra falls and pyroclastic flows. The significance of environmental hazards in the demise of civilizations such as Minoan Crete (tsunami) and the Early Bronze Age in the Near East (drought) has been vigorously debated for decades. While such events have undoubtedly threatened people in the region since prehistoric times, the actual threat to human society has increased dramatically in the historical and modern periods as urban environments and their populations have rapidly expanded. This part of the volume analyses hazards associated with both endogenic and exogenic Earth processes and the interactions between them. It includes volcanic processes, crustal instability, tsunamis, fluvial floods, extreme weather phenomena, and wildfires. Each chapter explores the basic controls and the geography of a particular hazard and related processes, and, over a range of timescales, magnitude and frequency relationships and the nature of the threat to human society.

High-magnitude events are a fundamental part of the physical geography of the Mediterranean and play a key role in long-term landscape evolution and ecosystem change. Even though the processes associated with each hazard typically take place over very short timescales, they can set in motion longterm adjustments to geomorphological and ecosystem processes. Tephra falls can change soil properties and vegetation communities, for example, and earthquakes may trigger base-level change and landslips in river basins that enhance fluvial sediment yields for many centuries. The Mediterranean has a distinctive meteorology that can produce extreme weather phenomena. A range of weather-related hazards has led to major losses of life on land and at sea. River basins in the region are characteristically small, with steep channels that have always been prone to high discharges. Yet even though flood magnitude decreased in some areas during the twentieth century, the flood hazard increased markedly as urban areas expanded across valley floors and along coastal plains. Wildfires devastated large areas of Greece in 2007 and climate change predictions for the twenty-first century suggest they will become an increasing problem across the Mediterranean. Other significant hazards are considered elsewhere in this volume and include storm surges and coastal flooding, mass movements, collapse features in karst terrains, dust storms, prolonged droughts, and heat waves. Each chapter in Part III examines hazard mitigation strategies and explores the problems that each hazard poses for Mediterranean society. Recent decades have seen the wider application of remote sensing and GIS tools for hazard zoning and increasingly effective collaboration between Mediterranean countries in systems of monitoring and warning.

This page intentionally left blank

15

Volcanoes Clive Oppenheimer and David Pyle

Introduction The historical record of Mediterranean volcanism is arguably the richest available for any region of the world. Documentary records date back to the Classical period, and archaeological records date back further still (Stothers and Rampino 1983; Chester et al. 2000). The Mediterranean is also home to some of the most famous, or indeed infamous, volcanoes on Earth (Table 15.1), several of which still present major threats to society today (Kilburn and McGuire 2001; Chester et al. 2002; Guest et al. 2003). A number, for example Santorini, Etna, and Vesuvius, have menaced human populations since Antiquity, and the human response and risk perception today are strongly shaped by a culture, which itself owes much to the volcanic landscapes and eruptions (Chester et al. 2008). The science of volcanology was born, and has since flourished, in the cradle of the Mediterranean. It began, arguably, with the careful descriptions by Pliny the Younger of the AD 79 eruption of Vesuvius that buried Pompeii, and developed through the scientific investigations of Sir William Hamilton in the eighteenth century. The region was the playground of the pioneers of modern volcanological studies in the nineteenth century (e.g. Fouqué 1879), and today it boasts a number of state of the art volcano observatories such as that which monitors Vesuvius. Several volcanoes, eruption styles, geothermal manifestations, and rock types have inspired nomenclature now widely used within the volcanological community: plinian, vulcanian, and strombolian eruptions; low temperature gas emanations known as solfataras; rocks known as pantellerites. The very word ‘volcano’ comes from the Aeolian island Vulcano, where Vulcan’s forge was situated. The sea-filled crater of Santorini was one of the

first volcanic ‘calderas’ to be described, by Ferdinand Fouqué in the 1870s.

General Geological Setting for Volcanic Activity The Mediterranean basin tracks the geological suture between the African plate to the south, and the Eurasian tectonic plate to the north (Chapters 1, 13, and 16). Many regions along this suture have experienced volcanic activity within the past 10–20 Myr (million years), most of it related to the continuing process of subduction that has consumed the northern margin of the African plate. Geologically, this has been a very messy collision, and our understanding of why particular volcanoes are found in particular places remains incomplete. During the past million years or so, the regions of active volcanism around the Mediterranean have become settled into the pattern that we see today (Figure 15.1). The active volcanic regions are more or less confined to central and southern Italy (from the Campanian and Roman provinces of the mainland, to the Aeolian Islands, to Etna, on Sicily, and Pantelleria; Figure 15.1); and to an arc of volcanism in the southern Aegean sea (from Aegina and Milos in the west, to Kos, Nisyros, and Yali in the east; Figure 15.1). Inland and further to the east, the trail of volcanism stretches through the volcanic province of Cappadocia, in Turkey, past the fabled, but poorly known, Ararat, and up into the great volcanoes of the Caucasus: Elbrus and Kazbek. In addition to the volcanoes that flank parts of the Mediterranean, mainland Europe has also seen geologically young episodes of volcanic activity, most recently as the ice sheets retreated, in regions including the

436

Clive Oppenheimer and David Pyle TABLE 15.1. Major ash layers correlated with known volcanic eruptions in the Mediterranean over the past 200 kyr

Eruption age (kyr BP)1 3.6 16.5 22.0 39.3 45 150 161 180 203

Eruption name

Other identifier (e.g. ash layer)

Source volcano

Composition

Magma volume (km3 )

Minoan Biancavilla-Montalto Cape Riva Campanian Ignimbrite Green Tuff Middle Pumice Kos Plateau Tuff Lower Pumice 2 Lower Pumice 1

Z2 Y1, Et1 Y2 Y5 Y6 W2 W3 V1 V3

Santorini Etna Santorini Phlegrean fields, Italy Pantelleria Santorini Kos Santorini Santorini

Rhyodacite Sodic andesite Rhyodacite Trachyte Peralkaline rhyolite Dacite Rhyolite Rhyodacite Rhyodacite

28–31

105–210 (>10) (>60) (>10) (>10)

Minimum tephra coverage area (km2 )

Additional references2

3 × 105 4 × 104 3 × 105 2–4 ×106 7 × 104

a b, c d e f

104 105 1.5 × 105

g

1

Eruption ages are ‘best estimates’ based on isotopic age constraints (calibrated 14 C or 40 Ar/39 Ar) or correlation with orbitally tuned marine core timescales. The correlation of V3 is newly suggested here. Additional references: (a) Pyle (1990); (b) Paterne et al. (1988); (c) Coltelli et al. (2000); (d) Wulf et al. (2002); (e) Pyle et al. (2006); (f) Civetta et al. 1988; (g) P. E. Smith et al. (1996). 2

Source: Compiled from data in Keller et al. 1978; Narcisi and Vezzoli 1999; Druitt et al. 1999; Lourens 2004.

Europe, which decompressed and melted during postglacial ice removal and flexuring (Nowell et al. 2006).

Auvergne and Massif Central of France; the Eifel region and the Rhine valley (Germany); around the flanks of the Pannonian basin of Hungary, Slovakia, and the Czech republic; and the Olot region of northern Spain (Scarth and Tanguy 2001). Geological explanations for the origin of these magmas are not agreed, but may relate to patterns of active rifting or to the presence of anomalously ‘warm’ regions of the mantle beneath

Volcanic Hazards Volcanoes can pose hazards between, as well as during, eruptions (Table 15.2). Volcanic disasters have often occurred months or years after eruptions begin, striking

Eifel Eifel

Chaine des Chaine desPuys Puys

Caucasus

Tuscany

Elbrus

Kazbek

Roman Province

Olot Olot

Campanian Province

Ararat

Amiata Colli Albani Ischia

Aeolian

Cappadocia Vesuvio

Stromboli i Lipari Vulcano Etna

Pantelleria Pntelleriaa

Campi Flegrei CampiFlegrei Mar di Sicilia

Kula Methana Methana Milos Santorini

Kos Kos Yali Nisyros

Aegean Volcanoes which have erupted since 1900 Other named volcanoes Other volcanoes

0

500 km

Fig. 15.1. Map of the Mediterranean basin showing the locations of selected volcanoes and volcanic provinces discussed in the text.

Volcanoes

437

TABLE 15.2. Summary of volcanic hazards Hazard type Pyroclastic density currents Debris avalanches, debris flows, lahars Tsunamis Airborne ash

Tephra falls

Landslides (debris avalanches) Lava flows

Point-source and diffuse gas and acid particle emissions (notably SO2 , sulphuric acid aerosol, HCl, HF, CO2 , H2 S, and radon)

Acid rain (due to rainfall through volcanic clouds) Ballistics (bombs, blocks) Earthquakes Ground deformation (e.g. subsidence, ground-cracking) Atmospheric shock waves Lightning in volcanic clouds Volcano restlessness or threat of eruption

Elements at risk, consequences Highly destructive to property and buildings, pyrogenic, cause thermal injury, and associated with a very low survival rate amongst exposed individuals. Drowning, impact injuries, extensive property destruction. Human health, property. Respiratory and cardiovascular health hazard (asthma, bronchitis, pneumoconiosis) and an irritant to eyes and skin, affects aircraft operations; impairs, disrupts, or prevents telecommunications. Damage to property, transport systems, human and animal health, water quality (corrosion of buildings; contamination of water supplies by volatiles, e.g. fluorine, carried on ash or blocking water treatment filters; impacts of agricultural/farming losses due to fluorine, burial under ash, etc.) Damage to health, property. Highly damaging to property, roads, agricultural land, forests, etc.; generally sluggish enough for people to evade except in some particular cases where lavas are very fluid and fast moving, or start uncontrolled fires. Impacts on air quality and human health, vegetation health and soil condition (agriculture). Acid gases cause bronchoconstriction, aggravation of respiratory disease; CO2 & H2 S: asphyxiation, possible neurological problems, nasal and lung cancers; radon creates lung cancer risk with long-term exposure. Chemical conjunctivitis and respiratory effects associated with HCl-rich clouds (known as LAZE for lava-haze) formed when lava enters the sea. Vegetation health (agriculture), property, water quality; irritant to eyes, skin. Impact injuries, burns. Property damage. Property damage and resulting physical injuries. Structural damage to property, roads. Property damage. Danger of electrocution. Economic impacts to regional development, tourism, investment, etc.

Source: Modified from Francis and Oppenheimer (2004) and Hansell and Oppenheimer (2004).

when the population at risk has become inured to the threat (Simkin et al. 2001). The intensity and magnitude of eruptions need not scale with their impacts, since exposure and vulnerability of elements at risk vary tremendously from one volcano to another. While the history of a volcano’s activity (witnessed in the rock record) points to the nature of potential future behaviour, the eruption style—and consequent hazards— are highly variable from one eruption to the next, and even within a single eruptive episode. Between eruptions, volcanoes may emit harmful gases and particles, and continue to pose serious hazards in the guise of mudflows, debris avalanches, and volcanogenic tsunamis. In a recent review of the literature, Witham (2005) found that 490 volcanic events in the twentieth century resulted in human impacts, with 4–6 million people evacuated, made homeless, or otherwise affected. Fatalities occurred in around half the events, with an estimated total death toll of 80,000–100,000. The risk of catastrophic human and economic losses from future eruptions is significant, especially given the spread of

urbanization in many volcanic regions—perhaps most notably in the Campanian district of Italy that includes the city of Naples. A further notable feature of the statistical record is that the number of injured (about 12,000) is much lower than the number of deaths (>80,000)— volcanic phenomena are often associated with low survival rates in impacted areas. Post-eruption famine and epidemics, pyroclastic density currents, mudflows, and volcanogenic tsunamis account for the majority of recorded deaths arising from volcanism. Few deaths have been associated with lava flows, however. While they may result in considerable damage to infrastructure and property they usually move sufficiently slowly to permit the evacuation of residents. The Mediterranean has escaped any major explosive activity since the early 1900s, when eruptions of Vesuvius in 1906 caused major economic impacts in Naples. One effect of these events was relocation of the 1908 Olympic Games, scheduled to take place in Rome, to London. Funds that were to have been used to support the Games in Rome were diverted to the reconstruction efforts in Naples. The following sections review a range

438

Clive Oppenheimer and David Pyle

(a)

(b)

Fig. 15.2. Volcanic hazards. (a) Oblique aerial view of Mt. Etna erupting in 2001. The dense ash cloud rises from the Laghetto cinder cone, formed during this eruption. (b) The 2001 eruption of Mt. Etna emplaced lava flows that destroyed part of the tourist village on the southern approach to the mountain. Fumes rise from the active margin of the flow. The tourist village has been protected in part by use of earth barriers but the road has been crossed by the new lavas (photos: Clive Oppenheimer).

of volcanic processes, products, and related hazards and are followed by a detailed examination of these phenomena from a Mediterranean perspective.

Tephra Clouds and Falls Ash clouds are associated with explosive eruptions (Figure 15.2a) and can pose a risk to aviation through damage to jet engines and abrasion of cockpit windows. To date there have been no reported air crashes arising from encounters with volcanic clouds, but there have been several near misses that resulted in substantial damage to the aircraft involved. Tephra fallout presents several kinds of hazard: by causing buildings to collapse due to loading on roofs and by impacts on terrestrial and aquatic environments. Heavy falls of tephra can damage vegetation, including agricultural crops, and pose a respiratory health hazard to humans. Tephra can also carry significant quantities of volatiles, including fluorine, scavenged from the eruption plume. Grazing animals can ingest toxic quantities of fluorine when ash contaminated in this way lies on the ground (Cronin et al. 2003). It may also lead to corruption of drinking water with the potential to cause fluorosis in human populations.

Pyroclastic Density Currents (PDCs) One of the most life-threatening aspects of volcanic activity results from exposure to clouds of hot gas and tephra known as pyroclastic density currents (PDCs; Branney and Kokelaar 2002). Past PDC events have resulted

in dead:injured ratios of 10:1 or greater—significantly higher than in any other type of natural disaster. PDCs typically form during the collapse of an explosive eruption column (e.g. Vesuvius in AD 79) or from gravitational failure of a lava dome (as witnessed on countless occasions at Soufrière Hills volcano, Montserrat, over the past decade). They can be envisaged as hot hurricanes of ash, rock, and gases, and can travel at speeds of 350 km hr−1 , or more, and reach temperatures of several hundred degrees Celsius. During an eruption of Mont Pelée on Martinique in 1902, PDCs swept into the town of St Pierre at an estimated speed of 160 km hr−1 , resulting in the deaths of 29,000 people within minutes (including all but two of the city’s population). The main causes of death are heat-induced shock, asphyxia, thermal lung injury, and burns. Survivors from PDCs tend to have been exposed to only the more dilute parts of the current or sheltered in some way (Loughlin et al. 2002), but can be critically injured due to respiratory and skin burns. Prior evacuation from areas at risk from PDCs is the only recommended way to minimize fatalities, for example through the establishment of exclusion zones. Many of the fatalities in Pompei and Herculaneum in AD 79 were believed to be from PDCs.

Lava Flows The term ‘lava flow’ usually refers to erupting lava that has the opportunity, and sufficiently low viscosity, to travel down the flanks of a volcano, or cross open

Volcanoes

ground. The expression is used both for active flows during emplacement, and for the resulting landform. In the course of a long-lived eruption, a lava flow field may develop by the superposition of many individual flow units. The current eruption of K¯ılauea, Hawai’i, has yielded over 2 km3 of lava and built up a flow field around 100 km2 in area since it began in 1983 (a mean eruption intensity of around 3 m3 s−1 ). Where lava enters the sea, as is the case in Hawai’i, a lava delta or bench may build seawards—though these are often unstable features. As active lava interacts with sea water on the shore in front of the lava bench, steam explosions can construct ephemeral littoral cones. Active lava flows radiate prodigious quantities of heat near the vent such that they rapidly form a surface crust. This can thicken sufficiently to insulate the core of the flow from thermal losses, lowering the rate of viscosity increase, and thereby promoting a longer travel distance. Mafic flows quite often crust over completely, with lava continuing to flow in tunnels, which can grow in cross-section by thermal erosion of the walls. When the supply of lava at the vent ceases, the last slug of lava may drain down-slope leaving an empty conduit or lava tube. On K¯ılauea, much of the lava flow between the Pu’u’ O’o vent, and the coastline where lava pours into the sea (a distance of over 10 km), takes place in a tunnel network, with only sporadic breakouts at the surface. Lava flows can result in dramatic destruction of property (Figure 15.2b), and loss of agricultural land and forest by burial and the action of fires.

Debris Avalanches, Debris Flows, and Lahars Because of their slopes and construction from sometimes poorly consolidated materials (rubbly lavas, loose ash), and through the weakening action of acidic groundwaters, volcanoes are prone to gravitational collapses. These may be triggered by magmatic activity within the edifice, seismic events, or heavy rainfalls. Even small volcanic landslides can be devastating in populated areas, and they can occur long after volcanic activity has ceased. This was the case with the landslide triggered by sustained and heavy rainfall at Casita volcano, Nicaragua, in 1998 (Kerle et al. 2003). In such cases, the landslides typically evolve into mudflows (see below). The major hazards posed by the various kinds of volcanic avalanche and flow phenomena are physical injuries related to burial and property damage, and drowning. Debris avalanches are fast moving, gravity-driven currents of partially or fully water-saturated debris—in

439

this context, primarily volcanic—that are not confined by an established channel. If the moving debris is watersaturated, and enters drainage channels, it is termed a debris flow, and if it consists of a significant fraction of clay-sized particles, it is called a mudflow (or lahar). Such flows can pick up water and debris along the way, while sedimenting out their coarser, denser materials, thereby transforming into clay- and water-rich hyperconcentrated flows. Lahars are a recurrent hazard at many volcanoes worldwide, and have claimed many thousands of lives. They form in a variety of ways, including the rapid melting of snow and ice by hot pyroclastic material, intense rainfall on loose volcanic deposits, and breakout of crater lakes, as well as being a consequence of debris avalanches. They can travel at speeds of 50 km hr−1 or more and travel many tens of kilometres. The lahar arising from the eruption of Nevado del Ruíz Colombia in 1985 covered more than 60 km and resulted in the deaths of an estimated 22,800 people.

Gas and Aerosol Emissions Volcanic gases and particles emitted into the atmosphere are ultimately deposited at the Earth’s surface, where they may have impacts on terrestrial and aquatic ecosystems, agriculture, infrastructure, and human health. The chemical and physical form in which they are deposited, and the spatial and temporal distribution of deposition, are strongly controlled by atmospheric chemistry and the transport of the volcanic plume. Various components of volcanic emissions (including acid species and heavy metals) can affect vegetation, and can have both harmful and beneficial effects. The detrimental effects are generally either mediated through acidification of soils (by dry or wet deposition) or by direct fumigation of foliage (e.g. respiration of acid gases through stomata). Chronic exposures to SO2 concentrations of a few 10s or 100s ppb are sufficient to affect plant ecosystems, decrease agricultural productivity, and cause foliar damage. However, other gas species (e.g., HF and HCl) can be important, as well as the extent of soil acidification due to wet and dry deposition (e.g. Delmelle et al. 2003). Several volcanic volatile species are harmful on contact with the skin, if taken into the lungs, or ingested (Figure 15.3; Hansell and Oppenheimer 2004). Few primary studies have been conducted into health effects of volcanic gases—those that exist are limited in terms of exposure assessment, so the true extent of health effects from volcanic gases is unclear. Most research to date relates to CO2 , H2 S, and SO2 exposures. Sulfur species

440

Clive Oppenheimer and David Pyle

Fig. 15.3. Fumarolic and diffuse soil emissions on Vulcano (Italy) pose a health hazard. Needless to say, this sign does not deter thousands of visitors from peering into the ‘smoke holes’ every summer (photo: Clive Oppenheimer).

(SO2 and H2 S gas, sulphate aerosol) can affect respiratory and cardiovascular health in humans (ibid.). Air quality in Hawai’i is reportedly affected by ‘vog’ (volcanic fog) associated with SO2 and sulfate aerosol from Kilauea’s plume, and ‘laze’ (lava haze), composed of HClrich droplets formed when active lavas enter the sea. More recent attention has focused on the abundance of very fine (sub-micron) and very acidic (pH about 1) aerosol emitted from volcanoes (A. G. Allen et al. 2002; R. S. Martin et al. 2008), which is likely to have human health consequences. Accumulations of H2 S in volcanic and geothermal areas, including faulty geothermal heating systems, have resulted in fatalities from asphyxiation. Communities in some geothermal areas are exposed chronically to low H2 S levels, notably in Rotorua, New Zealand, with possible impacts on the nervous system, and on respiratory and cardiovascular health.

Emissions of CO2 can also accumulate dangerously in low-lying areas and have resulted in deaths due to asphyxiation. Several cases have been reported at Furnas (Azores), Vulcano, Lazio, and Alban Hills (in Italy; see below), Cosigüina (Nicaragua), and Mammoth Mountain (California). Dissolved CO2 may also accumulate in lake water in volcanic areas. Sudden displacement of the water may release a cloud of CO2 able to flow under gravity, suffocating people and animals in its path, as occurred at Lakes Nyos and Monoun in Cameroon in the 1980s (Sigurdsson et al. 1987). A further volatile species encountered in volcanic and geothermal areas is radon (e.g. Avino et al. 1999). Many studies have considered an association between radon exposure and lung cancer in humans (e.g. Bowie and Bowie 1991; Avino et al. 1999; J. Pearce and Boyle 2005) raising some concern over the potential exposure in volcanic areas.

Volcanoes

441

TABLE 15.3. Twentieth-century record of deaths, injuries, and other impacts of volcanic activity in the Mediterranean Volcano

Phenomena

Date

Deaths

Injuries

Homeless

TOTAL

(inc. evacuees) Vesuvius Vesuvius Stromboli Santorini Etna Vesuvius Etna Stromboli Vesuvius Campi Flegrei Etna Etna Etna Campi Flegrei Etna Stromboli Etna Stromboli Stromboli Stromboli Stromboli Stromboli

T T(213)/L(3)/G(2) T T L L P T(24)/L(2)/I(1) S L T L S S T T T T T T T

10/03/1905 06/04/1906 22/05/1919 11/08/1925 07/11/1928 03/06/1929 02/08/1929 11/09/1930 00/03/1944 02/03/1970 04/08/1979 12/09/1979 17/03/1981 04/10/1983 16/10/1984 24/07/1986 17/04/1987 16/10/1993 01/06/1996 22/08/1996 04/09/1996 23/08/1999

1 218–700 4

1 300 20

100,000 10 houses many 4,300–5,000 60 houses

2 4–>6 22* 3 9

20 36,000 250 23–4 many houses

1 1 2

15+

250 <40,000

1,400

7 2 some 1 6 10

∗ Witham (2005) records 26–7 fatalities but the figure quoted here is considered more accurate (David Chester, personal communication, 2005). T = tephra falls and ballistics; L = lava flows; G = gas; S = seismicity; I = indirect causes; P = pyroclastic density currents.

Source: Witham (2005).

Volcanogenic Tsunamis On volcanic islands, landslides associated with eruptions can initiate a tsunami if the displaced material flows into the sea, or if the eruption itself is accompanied by collapse of an underwater edifice. In recent times, volcano-related tsunamis in the Mediterranean, for example, following eruptions of Colombos Bank (off Santorini in c.AD 1650) and Stromboli (2003) have been rather limited in extent and only locally destructive, with peak wave heights of the order of 1–2 m. Ancient volcanogenic tsunamis, such as those that accompanied the caldera-forming Minoan eruption of Santorini, are thought to have been highly destructive, as shown by the extensive sea-floor deposits of a strangely uniform sand deposit (‘homogenite’), and destruction along the northern coast of Crete (McCoy and Heiken 2000; Chapters 16 and 17).

Seismicity Seismicity associated with magmatic and volcanic activity can damage engineering structures. The seismic activity prior to large eruptions has left a clear archaeological record, certainly in terms of building structures damaged some time before eruption both at Pompeii

following the AD 79 eruption of Vesuvius, and at Akrotiri following the Minoan eruption of Santorini.

Active Volcanoes of the Mediterranean: Hazards, Crises, and Disasters The volcanoes and volcanic centres described here are selected on the basis of their inclusion in the Global Volcanism Program database (Siebert and Simkin 2002). It is by no means an exhaustive compendium of volcanism in the region since its focus is on volcanoes active since the beginning of the Holocene, though it also lists volcanoes noted for recent unrest (such as seismicity or hydrothermal discharges). The database nevertheless constitutes a convenient starting point, since these volcanoes may all be considered active or potentially active. Although in the last century there has not been a disaster in the Mediterranean on the scale of the AD 1631 or 79 eruptions of Vesuvius, active volcanism has continued to impact human populations in Italy and Greece in past decades (Table 15.3). Our review of volcanic hazards in the Mediterranean begins in northern Italy and ends close to Turkey.

442

Clive Oppenheimer and David Pyle

Tuscany Three volcanic centres are found in Tuscany: Larderello, Amiata, and Vulsini. Larderello, located in the northern Apennine Mountains, is the site of one of the world’s most productive geothermal power plants. A phreatic eruption was recorded in 1282 (Marinelli 1969), sited in the Lago Vecchiano crater. Amiata is composed of silicic lava domes and associated lava flows, and underlying ignimbrites, and is located about 20 km north-west of Lake Bolsena in southern Tuscany (Brogi 2008). Reaching 1,738 m a.s.l., Monte Amiata (La Vetta), is the largest of the domes. Radiometric ages indicate two main periods of activity at about 0.3 and 0.2 Myr ago (Ferrari et al. 1996). No Holocene products have been identified but Amiata also hosts ongoing hydrothermal activity notably at a geothermal field near Bagnore. Hot springs and gas vents emit mainly CO2 but also heavy metals, which are found concentrated in soils (Manasse and Viti 2007), lichens and mosses (Loppi and Bonini 2000). The Vulsini volcanic complex was formed as the result of several episodes of predominantly explosive volcanism occurring over the last 0.6 Myr. The products include tephra fall and pyroclastic density current deposits, some representing eruption magnitudes of several cubic kilometres (Karner et al. 2001; Palladino and Simei 2002). Major Pleistocene explosive eruptions resulted in the formation of the 16-km-wide, lakefilled Bolsena caldera and the 8 × 11 km Latera caldera (Palladino and Simei 2005). These events were followed by plinian eruptions in the Late Pleistocene. The last major eruption took place about 166 kyr ago at the north-west end of Latera caldera, and formed the Vepe caldera. There is, however, documentary evidence for a minor eruption in 104 BC, and the volcano remains seismically active (Amato et al. 1994).

Roman Volcanic Province The Roman Volcanic Province consists of several volcanic complexes, one of which is listed by the Global Volcanism Program: the Colli Albani (Alban Hills). They are located just south of Rome, and formed from a 600-kyrold caldera complex. It was thought that the last activity occurred around 30 kyr BP, forming a maar associated with the Peperino Albano ignimbrite until more recent overlying lahar and pyroclastic current deposits were discovered (Funiciello et al. 2003). Lake Albano is a maar that has overflowed catastrophically up to the Roman period. In 394 BC the Romans excavated a 1.5 km tunnel through the maar wall to maintain the lake 70 m below the low point on the maar rim. This drain, still in use, is

surely one of the earliest engineering structures for volcanic hazard mitigation (De Benedetti et al. 2008). Several other hydrovolcanic landforms are found in the area including Lake Nemi and the Valle Marciana, Quarto Laghetto, and Prata Porci craters. The eruptions that formed these features are thought to represent the final stages of the volcano’s life cycle and to be the result of small batches of magma rising into the water table (Funiciello et al. 2003). The area is heavily populated— the towns of Ciampino and Marino are sited within the complex. Surveys have been carried out in the Colli Albani to characterize potentially harmful emissions of CO2 , H2 S, and radon (Pizzino et al. 2002; Beaubien et al. 2003; Annunziatellis et al. 2003). Deaths of humans and animals in the region have been attributed to carbon dioxide asphyxiation (Beaubien et al. 2003). At one area, known as Cava dei Selci, CO2 emissions were found to contain up to 98 volume per cent CO2 and 0.8 to 2.0 volume per cent H2 S. The gas hazard near residential dwellings has been monitored by continuous measurement of CO2 and H2 S concentrations. Both H2 S and CO2 levels have exceeded environmental health limits, and prevention measures, including ventilation of basements, were recommended to minimize risk to residents (Carapezza et al. 2003). The high CO2 flux may result from the intrusion of magma into carbonate-rich crustal rocks (Funiciello et al. 2003). One plausible hazard scenario at Lake Albano, which is 2.5 km across at its maximum extent, is a sudden release of accumulated CO2 (Anzidei et al. 2008), as occurred at Lake Nyos in Cameroon in 1986 (Sigurdsson et al. 1987). Up to 200 mg l−1 of CO2 were found dissolved in the lake at a depth of 175 m (Martini et al. 1994), the highest concentrations found in Italian lakewaters. Rollover of the lake water could cause rapid exsolution and ‘syphoning’ of CO2 (Rice 2000). Eruptions within such lakes could also result in the release of volatiles dissolved in the water (e.g. Schmid et al. 2005).

Campanian Province Settled since Neolithic times, the Campanian region is considered one of the most high-risk volcanic regions on Earth. Its volcanoes include the Campi Flegrei (Phlegrean Fields) caldera complex, Mt. SommaVesuvio (Vesuvius), and the island of Ischia (Scandone et al. 1993; Civetta et al. 2004). They have impacted people and the environment since the earliest time of settlement (Cioni et al. 2000; Mastrolorenzo et al. 2002).

Volcanoes

443

Kostenki Kostenki

Likely extent of Y5 tephra fallout

Temnata Temnata

Minimum extent of Minoan (Z2) fallout

Campi Flegrei Minimum extent of Kos Plateau Tuff fallout

Santorini

Likely extent of Y5 tephra fallout

0

500 km

Fig. 15.4. Map to show the extent of fallout from the largest and most widely dispersed Mediterranean eruption of the past 200 kyr, the Y5/Campanian Ignimbrite eruption. Tephra fall deposits from this eruption, which originated from Campi Flegrei, have been found in lake beds and peat sequences in Greece, and in Palaeolithic archaeological contexts in Greece, Montenegro, Bulgaria (Temnata cave), and Russia (Kostenki). The easterly limit of ash deposition is not well constrained at the present time. For comparison, the extent of ash fallout during the 3.6 kyr Minoan eruption is also shown; again the limit of ash fallout downwind (to the east) is not well constrained. The extent of ash fallout from older eruptions (e.g. the 161 kyr Kos Plateau Tuff) is even less well known, due to lack of exposure, or preservation. Modified from Pyle (1990) and Pyle et al. (2006).

Today, around 3.5 million people live on or very close to these active volcanoes.

Campi Flegrei The Campanian Ignimbrite eruption produced some 100 km3 of dense magma 39.3 kyr ago, and initiated the formation of the Campi Flegrei, a complex of cones and craters that today lie within a 13 km

diameter caldera (Rosi et al. 1996). It was the largest eruption to have occurred in the Mediterranean for at least the last 200 kyr, and the associated ash deposits (known as the Y5 ash) were widely dispersed across the eastern Mediterranean and parts of eastern Europe (Figure 15.4). The extent of the ash fallout, covering an area in excess of 2 to 4 × 106 km2 with a recognizable deposit, makes the Y5/Campanian Ignimbrite event a

444

Clive Oppenheimer and David Pyle

key marker in both geological (it has been identified in many marine cores and in terrestrial deposits such as lake sediments) and archaeological contexts. Indeed, it is a critical marker bed in the Palaeolithic of Greece, Bulgaria, and south-western Russia, where its appearance coincides closely with a transition in stone tool technologies that, arguably, marks the arrival of Early Modern humans in Europe, and the beginning of the end for the Neanderthals (Fedele et al. 2002, 2003; Pyle et al. 2006). The Y5 tephra has been identified in Franchthi Cave in south-east Greece (Chapter 13). Russian archaeologists excavating the key Middle and Young Stone Age (Middle and Upper Paleolithic) sites at Kostenki, in the Don Valley, discovered this volcanic ash layer in situ in the late 1970s (Melekestsev et al. 1984); more recent excavations at Kostenki site 14 have found a cultural layer mixed in with the reworked tephra deposits, leading to speculation over the possible impacts of this colossal eruption on emerging societies (Sinitsyn 2003). Whether or not such impacts occurred, the real significance of the widespread Y5 ash layer will be as a time-marker, since the deposit has been exceptionally well dated at source (De Vivo et al. 2001; Ton-That et al. 2001), and the period of around 38–40 kyr BP is notoriously difficult to date with precision by radiocarbon methods, due to a large offset between ‘radiocarbon years’ and calendrical ages (e.g. Shackleton et al. 2004). The other major caldera-forming event linked to the Campi Flegrei was the eruption, about 16 kyr ago, of the Neapolitan Yellow Tuff (Deino et al. 2004), forming the yellow rock widely exposed around Naples and used in many of its buildings. Several periods of post-caldera volcanism have taken place, most recently on 29 September 1538 when the Monte Nuovo cone was born. The eruption culminated in an explosion on 6 October that killed twenty-four people. Today, almost 1.5 million people live within the caldera. The caldera has also exhibited remarkable linked seismicity and ground deformation, focused historically around the town of Pozzuoli. A new bradyseismic crisis began in 1982, possibly triggered by a Magnitude 6.9 earthquake centred 100 km to the south-east (Orsi et al. 1999). Earthquake damage rendered many buildings unsafe, and continuing swarms of tremors caused widespread concern. Between June 1982 and December 1984, 1.8 m of uplift were measured at Pozzuoli, the peak rate of uplift reaching 5 mm per day during periods of high seismicity in September and October 1983 and March and April 1984. The cumulative net uplift since 1969 was 3.5 m. Seismicity decreased to low levels in 1984, and some subsidence took place, at a rate of 0.4 mm per day. The caldera hosts an important

magmatic-hydrothermal system, notably manifested at Solfatara near Pozzuoli. Analysis of data acquired from a high resolution seismic reflection study indicates the presence of a very large mid-crustal magma body beneath the caldera (Zollo et al. 2008).

Vesuvius Few volcanoes can match Vesuvius in terms of the richness of the documentary record of volcanism. Its written record stretches back famously to Pliny the Younger’s account of the AD 79 eruption (Figure 15.5). Vesuvius erupted on many occasions in the Mediaeval period and up to the last event in 1944 (Stothers and Rampino 1983; Principe et al. 2004). It is one of the deadliest volcanoes on record, having claimed at least 8,000 lives (Table 15.4; Figure 15.6). The civil protection authorities and volcanologists working on volcanic risk arising from future activity of Vesuvius have focused on a reference event, believed to represent the maximum expected eruption. This hypothetical scenario is based on the subplinian eruption of 1631. Precursors are likely to include seismicity, some of which could be damaging. Festa et al. (2004) have modelled pre-eruptive seismicity for Vesuvius and its effects on the Campanian region. They evaluated strongmotion parameters, which are useful for estimating the seismic damage to the built environment that would arise from both the maximum expected single event and the cumulative effect of a large number of small events. The scenario eruption itself is envisaged to begin with phreatomagmatic explosions, opening up the conduit of the volcano, and associated with strong earthquakes. Once the conduit is open, a sustained eruptive column could ascend to 15–20 km above sea level. Tephra fallout would probably result in extensive building damage

TABLE 15.4. Fatal eruptions of Somma-Vesuvio Year (AD) 79 1631 1682 1737 1794 1805 1872 1906 1944

Number of fatalities 3,500 4,000 4 2 400 4 9 350 27

Cause P, T P T

L, G T T, L

Note: P = pyroclastic density currents; T = tephra falls and ballistics; L = lava flows; G = gas. Sources: Simkin and Siebert 1994; Tanguy et al. 1998.

Volcanoes

445

Fig. 15.5. View of Herculaneum and modern Ercolano. The Roman town was buried by up to 20 m of tephra from pyroclastic density currents during the AD 79 eruption. The modern town (Ercolano) has grown up directly on top of the deposits, and multistorey buildings crowd the edge of the excavation (background). Monte Vesuvius broods in the middle distance. Note also the row of chambers at the lower right of the photograph, which lined the shoreline at the time of the eruption. Hundreds of skeletons of victims were discovered inside them (photo: Clive Oppenheimer).

(Barberi et al. 1990; Lirer et al. 2001a; Cioni et al. 2003). As occurred during the 1631 event, the eruption column is likely to founder at some point, generating PDCs that could affect an area of up to 230 km2 (Dobran et al. 1994; Todesco et al. 2002; Esposti Ongaro et al. 2002). These currents will be highly destructive and threaten the lives of all caught in or near them. Skeletons of victims found in boat sheds at Herculaneum (killed during the eruption of Vesuvius in AD 79) suggest brief (a few minutes) exposure to temperatures of about 500◦ C (Mastrolorenzo et al. 2001). Nunziante et al. (2003) and Petrazzuoli and Zuccaro (2004) have considered the impacts of such pyroclastic currents on the building stock. A further potential consequence is damage to cultural heritage sites such as Herculaneum, Pompeii, and Villa Oplonti for which some imaginative solutions have been proposed by Patella and Mauriello (1999). Of course, there will be wider impacts of future activity, including major disruption to aviation (Macedonio et al. 1994). Structural failure of the volcanic edifice could also occur, triggering landslides and potentially caldera formation. In the waning stages of an eruption, magma-water interaction in the conduit could produce phreatomagmatic explosions, while rainfall would redistribute the tephra fallout, generating destructive lahars. The potential impacts of larger eruptions, such as an AD 79 scale event, have also been explored (Lirer

et al. 1997). Analysis of the complex pattern of longterm poissonian and shorter term clustering of historic eruptions at Vesuvius has led to one prediction that the next subplinian eruption is likely to occur around the year 2030 (Palumbo 1997). Despite the lack of volcanic activity in recent decades, the historic tephra fall deposits from Vesuvius are remobilized occasionally, as occurred with devastating results in May 1998 in the Sarno region (Pareschi et al. 2000b; Toyos et al. 2003; Zanchetta et al. 2004). The Sarno area is on the margin of the Apennine belt, some 20 km east of Somma-Vesuvio, and bordered to the south-west by the Campanian Plain graben. The massif, which is composed of Mesozoic-Cenozoic carbonate rocks, is mantled with volcaniclastic deposits derived mainly from explosive activity of Somma-Vesuvio (Pareschi et al. 2002; Frattini et al. 2004). On 5 and 6 May 1998, immediately following intense rains and a particularly wet spring, slope failures generated several hundred landslides that coalesced and rapidly transformed into debris flows. These inundated several areas including the towns of Quindici, Siano, and Bracigliano, and the Clanio valley. There were more than 150 fatalities and extensive damage to property. In addition, Catapano et al. (2001) documented adverse mental health consequences in the population affected by the 1998 disaster. Evidence for much larger flank failures has recently been identified in high-resolution offshore seismic profiles in the Bay of Naples. Milia et al. (2003)

446

Clive Oppenheimer and David Pyle

Fig. 15.6. Plaster cast of one of Pompeii’s victim (photo: Clive Oppenheimer).

recognized two debris avalanche deposits, one with a volume of about 2.9 km3 , and a stratigraphically higher deposit with a volume of about 1 km3 . They tentatively linked these to two major Plinian eruptions of Somma-Vesuvio, namely the 18 kyr BP Pomici di Base and 3.4 kyr BP Avellino eruptions. They estimated that the debris avalanches initiated subaerially and travelled about 12 km before reaching the sea. These events raise implications for tsunami hazard associated with any future large-scale flank failure of Somma-Vesuvio.

Ischia The oldest volcanic rocks on the island of Ischia have not been dated, but part of the more recent stratigraphy (75–50 kyr BP) has recently been characterized in detail (Brown et al. 2008). This study reveals a rich volcanic history, which includes the eruption of the Mt. Epomeo Green Tuff, a caldera-forming event about 55 kyr ago. The most recent phase of activity began around 10 kyr

BP , centred in the east of the island and involving lava flows and phreatomagmatic and magmatic explosions. Similar activity has recurred during the Holocene, and the most recent recorded eruption took place in 1302 AD (Capaldi et al. 1976/7). This eruption formed a spatter cone and emplaced a lava flow as far as the north-east coast. Mt. Epomeo has experienced at least 800 m of resurgent uplift associated with seismicity and gravitational instability (Poli et al. 1989). Large flank failures occurred from 8.6 to 5.7 kyr BP (Tibaldi and Vezzoli 2004), raising concerns over tsunami risk. Alberico et al. (2008) have also evaluated the hazard of pyroclastic density currents on the island. Fault systems intersect the island, and have generated disastrous earthquakes in the historic period, notably in 1228, and then in 1883, when the town of Casamicciola on the north of the island was destroyed with the estimated loss of 2,333 lives (Cubellis et al. 2004). The latter event was the first major disaster to befall Italy following its unification in 1860, and it led to the passage of the first legal act concerning seismic safety in the nation (Chapter 16). In addition, there are many fumaroles and thermal springs on the island (which have been used since Roman times as thermal baths), associated with a mixed hydrothermal reservoir of magmatic, meteoric, and sea-water fluids (Chiodini et al. 2004). The strongest fumarolic activity is concentrated on the south-western flanks of Mt. Epomeo. In several locations, the groundwater contains high concentrations of several elements, notably arsenic, which is found at levels exceeding health limits (Lima et al. 2003). Today, the island population stands at around 50,000, though numbers swell during the tourist season.

Aeolian Islands The Aeolian Islands are located in the Tyrhennian Sea between Sicily and mainland Italy. The volcanism began in the Pleistocene, and has built up a 200-km-long arc composed of seven subaerial volcanoes (Alicudi, Filicudi, Salina, Lipari, Vulcano, Panarea, and Stromboli) and several seamounts that roughly surround the Marsili basin (Beccaluva et al. 1985). The explosive nature of volcanic activity that characterizes the eruptive style of the islands, and their tsunamigenic potential (Maramai et al. 2005a ) pose significant short- and long-range hazards.

Panarea Panarea is the smallest of the Aeolian Islands at just 3 by 1.5 km, though it is nevertheless a significant volcanic edifice rising about 2,000 m above the sea floor.

Volcanoes

447

Fig. 15.7. Sequence of laminated deposits from pyroclastic density currents in the Monte Guardia area of Lipari. The darker layers are interpreted to be the result of water-rich explosive eruptions. They show characteristic ‘bomb-sags’ where volcanic projectiles have thwacked into the soft sediment (photo: Clive Oppenheimer).

Stratigraphic and volcanological evidence indicates that Panarea was formed from initial effusive and later explosive activity mostly during a relatively short time period 150 to 125 kyr ago (Calanchi et al. 2002). The youngest volcanic products found on the island are tephra falls dated to between 42,000 and 13,000 yr BP , though these may have originated elsewhere since they are also found on the other Aeolian islands. The main volcanological interest today is the submarine fumarole field found in shallow waters offshore. In November 2002, a sudden increase in degassing was observed (Capaccioni et al. 2007), possibly following a submarine explosion (Caliro et al. 2004), which appears to have been triggered by an input of magmatic volatiles. The event was accompanied by a low intensity seismic swarm, and resulted in the forma-

tion of a small crater on the seabed. Residents on the island reported smelling H2 S. After a few days, the degassing subsided.

Lipari Lipari is the largest of the Aeolian islands, and is built up from small stratovolcanoes, craters, and lava domes on a basement of submarine volcanic deposits. The latest eruptive cycle, which began around 40 kyr ago, produced rhyolitic lavas and tephra and includes the Mt. Guardia pyroclastic current deposits emplaced 22–24 kyr BP (Figure 15.7; De Rosa et al. 2003). The most recent eruption formed two impressive obsidian lava flows: the Rocche Rosse and Forgia Vecchia. These have been dated to either the sixth (Crisci et al. 1991) or eighth centuries AD (Cortese et al. 1986). An associated

448

Clive Oppenheimer and David Pyle

tephra has been identified in archaeological context in Southern Albania (Bescoby et al. 2008).

Stromboli Stromboli is the northernmost of the Aeolian islands, and is situated some 60 km from the Calabrian coast. It rises from 1,500–2,000 m below sea level to an altitude of 924 m (Figure 15.8a). Stromboli has been active for as long as records exist, and its characteristic style of periodic, low-intensity eruption is recognized at many other volcanoes where it is referred to as strombolian activity (Figure 15.8b). In the last 13 kyr, Stromboli has experienced at least four major collapses of its north-west flank involving avalanche volumes of 1–3 km3 (Apuani et al. 2005). The youngest has formed the steep Sciara del Fuoco down which tephra and lava flows have plummeted to the sea in the modern period. The typical activity of Stromboli today involves explosions every 10–20 minutes or so that expel lava bombs, lapilli, and ash from

magmatic vents located within the crater terrace. The explosions result from the rupture of gas bubbles at the top of the magma-filled conduits, and last a few seconds. In addition, there is a steady stream of magmatic gases into the atmosphere. Larger events occur occasionally, producing more violent explosions that have in the past killed and injured tourists, and issuing lava flows that descend the Sciara del Fuoco. There are two strips of coastline that are settled and these have been affected by stronger explosions every ten years or so (Barberi et al. 1993). Slope failures on the Sciara del Fuoco have in the past generated destructive tsunamis, most recently in December 2002 (Bonaccorso et al. 2003; Maramai et al. 2005b; Chiocci et al. 2008) following a phase of strong activity that began in May of that year. This event began as a submarine landslide but led to detachment of subaerial material at a height of 650 m above sea level. Run-up heights of several metres were documented on Stromboli’s coast, associated with considerable building damage. Tinti et al. (2003) modelled tsunamis originating at Stromboli and concluded that a major failure of the north-western flank of the cone is capable of generating a regional tsunami that would affect long stretches of the Tyrrhenian coastline, but especially the neighbouring islands of Panarea and Salina, and parts of the Calabria coast (Chapter 17).

Vulcano The eponymous volcano has been vigorously active through much of the Classical and historic period (Arrighi et al. 2006; Figure 15.9). About 6 kyr ago, activity became focused in the centre of the Fossa

Fig. 15.8. Stromboli volcano. (a) Stromboli is a substantial volcanic edifice. Some two-thirds of its 3 km height is below sea level. The small stack to the right of the photograph (Scoglio Spinazzola) is a remnant of a flank cone on the submarine part of Panarea island. (b) Time-lapse photograph of a typical strombolian eruption on Stromboli (photos: Clive Oppenheimer).

Fig. 15.9. Vulcano seen from Lipari. The closest part of the island seen in the photograph is the young volcanic centre, Vulcanello, which rose from the sea in the second century BC. La Fossa looms behind it (photo: Clive Oppenheimer).

Volcanoes

caldera, leading to the formation of the still-active Fossa cone. More recently, a new eruptive centre, Vulcanello, has grown in the strait between Vulcano and Lipari. Its first recorded eruption took place in 183 or 123 BC forming a new island. Sporadic eruptions continued here until the mid-sixteenth century AD. By that time, the new island had significantly enlarged and eventually connected with the main island of Vulcano. Vulcanello has not erupted for 400 years, and fumarolic activity declined after the mid-nineteenth century. La Fossa’s last eruption in 1888–90 was carefully described by Mercalli and Silvestri (1891). Since the 1920s, the cone has hosted a vigorous magmatic-hydrothermal system, one of the most carefully studied in the world. Because of the explosive nature of activity and the proximity of the main village on the island to La Fossa, there have been considerable efforts to understand the nature of the volcanic hazard (e.g. Frazzetta and La Volpe 1991; Montalto 1995) and the complexities of Vulcano’s plumbing system. Although the permanent population of the island numbers only around 500 (many of which inhabit the Piano region a few kilometres from La Fossa), at the height of the tourist season, around 10,000 people can be found within a few hundred metres of the crater. Vulcano is also prone to flank failure. A minor landslide generated a tsunami in 1988 (Tinti et al. 1999). In addition to the strong fumarolic degassing around the Gran Cratere of La Fossa, there are numerous sites of diffuse degassing on the flanks of the cone and along the shore. These represent a CO2 hazard, as found in the Colli Albani. Several deaths have been attributed to asphyxia (Baubron et al. 1990), prompting efforts to model the

449

gas dispersion and associated human health hazard (e.g. Pareschi et al. 2001). Klose (2007) has evaluated health hazard associated with SO2 emissions on the island.

Sicily and Pantelleria and Submarine Volcanoes Etna Mount Etna, located in eastern Sicily, is one of the most frequently active volcanoes on Earth (Salvi et al. 2006). Its eruptions have been documented since the Classical period. Although its historical eruptions have been predominantly effusive in character, they are often associated with strombolian activity, fire fountaining, and sustained ash cloud generation, and there have been much larger explosive phases (Chester et al. 1985) notably in 122 BC (Coltelli et al. 1998). Sicily’s largest city, Catania, is located on the south-eastern foot of Mt. Etna (Figure 15.10), and has been reached by lava flows from Etna in the past. The volcano presents multiple hazards to the surrounding communities both from summit and flank activity (e.g. Guest and Murray 1979; Dibben 2008) and shows apparently clear temporal patterns in its activity that may assist in long-term eruption forecasting (e.g. Palumbo 1998; Murray 2003). Arguably the most significant eruptions of Etna during the historic period were the 1381 eruption, in which lava reached the north-eastern part of Catania, and the 1669 eruption during which lava flows reached the sea. The 1669 eruption began on 17 March, with lava emerging from numerous vents near Monte Rossi at 750 m above sea level on the south-east flank of the volcano. Several towns, including Malpasso, Mascalucia,

Fig. 15.10. Aerial view of Mt. Etna rising above the city of Catania. Etna’s gas and aerosol plume can just be made out, rising above the summit of the volcano (photo: Clive Oppenheimer).

450

Clive Oppenheimer and David Pyle

Camporotondo, S. Pietro Clarenza, S. Giovanni Galermo, and Misterbianco, were affected. On 1 April the flows reached Catania and, on 23 April, the coast, where they formed a 2-km-wide flow front. Over the four-monthlong eruption, 1 km3 of lava was emplaced (an average effusion rate of about 100 m3 s−1 ) generating a flow field 17.5 km long covering an area of 37.5 km2 (Crisci et al. 2003). When the flows reached Catania, the city walls initially deflected them but were soon breached, allowing small lobes of the main flow to cause some damage to buildings. Only one other historic eruption has generated so much lava, that of 1614–24, though the effusion rate was much lower for this earlier event. Several other eruptions have damaged property, notably the 1928 eruption, which caused destruction in the town of Mascali (Duncan et al. 1996). The ash production during Etna eruptions has caused difficulties for aviation, with Catania’s international airport, Fontanarossa, and associated airspace, occasionally affected by clouds and falls of tephra (Figure 15.2a). Some deposits from ancient eruptions of Etna contain a fibrous amphibole mineral linked with a cluster of mesothelioma cases (Biggeri et al. 2004) raising some concern about the potential health hazard of Etna ash, though no fibrous habit has been reported to date in any recent tephra falls. Etna’s persistent gas and acid aerosol plume is one of the most prolific of any volcano, emitting, on average, several thousand tonnes of SO2 into the atmosphere every day (Allard et al. 1991), along with halogens and other components (e.g. Francis et al. 1998). Recent observations indicate that the plume can reach the ground 10 km or more from the summit (A. G. Allen et al. 2005). Chemical burning of leaves and flowers of vegetation downwind of Etna has been reported (Cimino and Toscano 1998), as well as some limited impacts associated with the deposition of fluorine (Notcutt and Davies 1989; Bellomo et al. 2003) and trace metals (e.g. Grasso et al. 1999). Mount Etna exhibits a range of ground deformation phenomena—not all of which are related to seismicity— and during the Middle and Late Pleistocene these have impacted upon the development of river systems that drain the slopes of the volcano (Chapter 11). At the largest scale, the Valle del Bove, which cuts the eastern flank of the volcano, represents the outcome of major flank failure (e.g., Calvari et al. 2004).

Pantelleria Pantelleria is an island, some 15 km across, situated in the Strait of Sicily. It hosts two large Pleistocene calderas,

which enclose various lava domes and cinder cones. The 6-km-wide Cinque Denti caldera is the younger of the two and formed about 45 kyr ago (Orsi et al. 1991). A submarine eruption in 1891 from a vent off the north-west coast is the only confirmed historical activity (Siebert and Simkin 2002). Geodetic measurements have detected subsidence of the Cinque Denti caldera, possibly indicating ongoing adjustment to prehistoric eruptions (Bonaccorso and Mattia 2000). The volcanic complex also hosts a vigorous magmatic-hydrothermal system. Geodetic studies have revealed recent subsidence of the caldera but it is difficult to discriminate between hydrothermal and magmatic origins of these displacements (Mattia et al. 2007).

Campi Flegrei Mar Sicilia A curious eruption took place in 1831, 50 km northeast of Pantelleria, and Charles Lyell described the event in his Principles of Geology. The new island was immediately claimed for Britain, Spain, and Italy. Each nation proposed a different name for it but the island fairly soon succumbed to wave action and disappeared. The volcano appeared to be stirring again in 2002, as Italian seismologists reported new earthquake activity in the vicinity. Such submarine activity can pose a severe threat to shipping.

The Aegean Province The volcanoes of the Aegean form a textbook ‘arc’, whose existence is linked to the subduction of the African plate beneath Crete. Over the past few million years, the floor of the Aegean has been spreading southwards to such an extent that a chain of extinct volcanoes, which formed the active arc just 3–6 Myr ago, now lies stranded many tens of kilometres to the north of the present arc (Pe-Piper and Piper 2002).

Methana Methana, a small rocky peninsula that emerges from the Peleponnese into the Saronic Gulf, south of Athens, comprises a sequence of young volcanic rocks built upon a limestone basement over the past million years or so. Little detail is known of the volcanic history of this peninsula, but it is believed that the last eruptive activity—the extrusion of an andesite lava flow and minor associated tephra at Kameno Vouno—took place between 276 and 239 BC (Stothers and Rampino 1983). Today, and for most of the historical record, Methana remains better known for its hot springs (on land), and occurences of strongly metal- and metalloid-enriched

Volcanoes

sediments offshore (e.g. Hübner et al. 2004), which point to continuing submarine hydrothermal activity.

Milos Milos is one of the larger islands of the Greek Cyclades, covering an area of 160 km2 . While the last known eruptions occurred in Roman times, it remains one of the more prolific sites of submarine hydrothermal activity in the Aegean. Most of the activity of the past 50,000 years has probably taken the form of phreatic explosions. The subaerial part of the island covers only about 20 per cent of the full extent of the volcanic edifice, according to interpretation of seismic profiles (Anastasakis and Piper 2005). This edifice began to grow about 5 Myr ago, perhaps related to the onset of rapid lithospheric extension in this part of the Aegean (van Hinsbergen et al. 2004), and was built up throughout the Pliocene from submarine and shallow intrusive domes and tephra deposits, forming an extensive complex. Several of these submarine domes are now exposed above sea level, and their importance for the recognition and understanding of shallow submarine volcanism is becoming clear (e.g. Rinaldi and Venuti 2003; Stewart and McPhie 2003). These domes, some of which are of obsidian, have been an important economic resource since the Bronze Age. Obsidian from Milos has been recovered from the Mesolithic deposits at Franchthi Cave in southeast Greece (Chapter 13), thus providing some of the earliest evidence of seafaring in the Mediterranean. The

451

volcanic domes on Milos continue to be mined today, both for their primary and alteration products. Altered volcanic deposits on Milos host the only young gold deposits known from the Aegean (Kilias et al. 2001).

Santorini Santorini is the best-known and most prolifically active volcano of the South Aegean volcanic arc. It is also one of the most readily identifiable volcanoes of the world, by virtue of its distinctive plan form: with a ring of islands (Thera,Therasia, and Aspronisi) defining the outer edge of a complex flooded caldera, and the two Kameni (literally: ‘burnt’) islands emerging from the centre of the caldera. Santorini has a rich volcanic history, exquisitely preserved in the steep caldera cliffs (Figure 15.11). More than twelve major explosive eruptions are recognized, spanning the past 250 kyr (Druitt et al. 1989, 1999), at least four of which were accompanied by major episodes of caldera formation. Many of these pyroclastic eruptions were followed by extended periods of shield building, leading to the intercalations of lava shields and tephra that are exposed in the caldera margins. Currently, Santorini volcano is in a shield-building phase, with the focus of intra-caldera volcanism for the past 2,200 years being the Kameni islands (Figures 15.12 and 15.13). These islands, of which there are currently two (Palea and Nea Kameni), have been subject to extensive investigation, on account of their unusually uniform chemical compositions, their

Fig. 15.11. The town of Fira clinging to the rim of Santorini’s caldera (photo: Clive Oppenheimer).

452

Clive Oppenheimer and David Pyle

Fig. 15.12. Map of part of Santorini in c.1715, showing the Kameni islands after the 1707 eruption. The Grande isle Brulée (Grande Cameni) is now known as Palea Kameni; the Île Nouvelle, formed during the 1707–12 eruption, now forms a part of Nea Kameni, as does Petit Cameni (or Mikra Kameni). Modified from D’Armenonville (1715), published by permission of the Syndics of Cambridge University Library.

heterogeneous population of exotic inclusions, and their remarkably well-documented eruptions of the nineteenth and twentieth centuries (e.g. Nicholls 1971; Pyle and Elliott 2006). There are considerable ongoing efforts to characterize and monitor the state of the Kameni islands, in particular their seismicity (Dimitriadis et al. 2005), hydrothermal activity (P. A. Smith and Cronan 1983), and geodetic motion (Stiros and Chasapis 2003). Although the islands are currently in repose, there is no reason to suppose that there will not be future eruptions of a similar nature, perhaps within decades, and almost certainly within centuries. But Santorini is best known for its last, large explosive eruption: the Bronze Age ‘Minoan’ eruption, which

took place about 3,600 years ago. The social and environmental impacts of this catastrophic caldera-forming eruption first came into focus in the 1930s, when Spiridon Marinatos began an extensive series of excavations of a Minoan port town at Akrotiri, on the southern flank of Santorini (Marinatos 1939). ‘Akrotiri’ (its ancient name is unknown) had been a prosperous town, which was subsequently damaged in an earthquake, perhaps some years prior to the eruption, and then abandoned before being buried by the tephra fall and pyroclastic current deposits of the Minoan eruption. The scale of the event, in which nearly 30 km3 of rhyodacite magma were discharged, led to speculation that the eruption had triggered social upheaval not only on Santorini (which

Volcanoes

453

Fig. 15.13. (a) Map of Nea Kameni and Mikra Kameni, after the 1866–70 eruptions, the lava from which erupted mainly from the dome and crater complex in the centre of the island. Note that the 1707–12 crater vent still survives at this point. Modified from Fouqué (1879). (b) Hillshade digital elevation model of the present-day status of the Kameni islands, based on airborne laser-ranging measurements in 2004. Nea Kameni was substantially modified during the eruptions of 1925–8, at which point the cone of Mikra Kameni (c.AD 1570) was encircled by lava, and during eruptions in 1939–41. The crater from 1707–12 was finally destroyed during these latter eruptions. Note that many of the lava flows take the form of channels with steep sides (levées). The abundant surface textures are wrinkle folds, formed during lava emplacement. Modified from Pyle and Elliott (2006).

454

Clive Oppenheimer and David Pyle

was completely covered by metres of tephra during the eruption), but also on Crete and across the Aegean (the extent of ash fallout during the eruption is shown in Figure 15.4). Santorini also became popularly linked to the mythology of Atlantis (e.g. Luce 1969). The possible implications of the event spawned a series of interdisciplinary conferences in 1978 and 1989 (Doumas 1978; Hardy 1990), with the hope, amongst others, that the precise dating of the eruption might unify discordant eastern Mediterranean archaeological chronologies and provide natural scientists with a better understanding of the potential global effects of large explosive eruptions (Manning 1999). Despite much endeavour—and substantial progress in understanding the physical volcanology of the eruption (e.g. Bond and Sparks 1976; Sparks and Wilson 1990)— some of the chronological issues remain unresolved, in the main because of the vagaries of atmospheric 14 C production around 3,500 to 3,700 years BP. The variability in 14 C production rate over this period has led to a plateau in the calibration curve, such that with radiocarbon alone it is not possible to derive as precise an eruption age as one might wish (e.g. Manning et al. 2002). Attempts to link the eruption to specific sulphate peaks in Greenland ice cores, and to tree-ring anomalies (e.g. Hammer et al. 1987; Baillie and Munro 1988) have been called into question (N. J. G. Pearce et al. 2004), and it now appears that the eruption probably took place between 1599 and 1663 BC (Ramsey et al. 2004), and possibly between 1600 and 1627 BC (Friedrich et al. 2006). The Minoan event was the last major silicic eruption of Santorini, and possibly the final phase of the second of two cycles of explosive activity on this island: both cycles beginning with large andesitic-to-dacitic explosive eruptions, and terminating with a pair of large rhyodacite eruptions (Druitt et al. 1989). Assuming that the pattern of future explosive activity on Santorini follows a similar path, the next major explosive eruption may be of andesite and will have a vent location close to the present-day Kameni islands. However, there is a less than 50 per cent likelihood of such an event within the next 10,000–20,000 years (Pyle and Elliott 2006). Instead, it is considerably more likely that there will be further intra-caldera activity at the Kameni islands within the next century. Historical activity of the Kameni Islands has followed a typical pattern of behaviour. Unrest has begun with uplift of regions of the sea floor, indicating shallow magma intrusion; leading to increasingly vigorous hydrothermal activity in the days before the eruption.

Eruptions have usually started relatively calmly, with the rapid growth of a dacite lava dome, but have progressed to intermittent explosive (vulcanian) activity at the dome accompanied by sustained lava effusion. During the eruptions of 1866 and 1939, the dome growth history was essentially identical, with the domes reaching heights of 20 m after 3 days; 50 m after 5–7 weeks, and growth ceasing after 3–4 months. In recent years, the main phases of activity have lasted 1–3 years, with activity gradually declining after an early peak, the length of the eruption apparently relating to the period of pre-eruptive repose (Figure 15.14). The lavas extruded during these eruptions are typically a blocky ¯ which creeps slowly out to a distance of 500– dacite ‘a‘a, 1,000 m and forms distinctive, broad lava lobes with well-defined marginal levees and an axial array of en echelon ridges (Figure 15.13b). All these features are explicable in terms of the high viscosity of the magma (Pyle and Elliott 2006). The well-documented patterns of historical behaviour at the Kameni islands provide good indications to the likelihood and nature of future activity (Fytikas et al. 1990). It is clear that the major hazards will be acidic gas emissions and the vulcanian explosions. The former were documented in the first major study of the medical effects of volcanic emissions following the 1866 eruption by da Cologna (1867). Recent work provides little comfort, though, in terms of the length of any preeruption warning at the Kameni islands. All of the recent lava flows carry partially crystalline inclusions, which are thought to be remants of the magma whose intrusion triggered the eruption (V. M. Martin et al. 2006). Zoning patterns of minerals in these inclusions point to an interval of weeks or less between the intrusion event, and eruption (V. M. Martin et al. 2008). In contrast to the activity of the Kameni islands— where there is a high probability of future eruptions, but whose nature will be of comparatively minor impact— the AD 1650 eruption of Columbos Reef (also known as Colombos bank) is an example of the style of eruption that poses the greatest concern to scientists working to mitigate the impacts of future activity on Santorini. Columbos Reef is a submarine volcano that lies about 6 km north-east of the island of Thera. Its 1650 eruption was well documented by Jesuit missionaries at the time, and many translations of the original reports have since been published (e.g. Pègues 1842; Leycester 1851; Fouqué 1879). The activity was presaged by six months of seismicity, several shocks being ‘so frightful and furious that the houses rocked to and fro’. The eruption began on 27 September 1650, with considerable

Volcanoes

Eruption length (days)

2000

(a) 1707

1500 1000

1925 1939

500

1950

0

100 150 50 Inter-eruption interval (years)

0

1000 Dome height (metres)

1866

(b)

200

Kameni 1866, 1939 domes St. Vincent 1979 St. Helens 1981–1986

455

reportedly inundated about 2 km2 of land along the east coast of Santorini, causing damage to buildings and agricultural land. However, although nineteenthcentury Admiralty charts refer to the locations of submerged ruins offshore from present-day Kamari (Leycester 1851), recent investigations on land failed to reveal any strong evidence for the impact of the tsunami (Dominey-Howes et al. 2000). A recurrence of the AD 1650 event, were it as described in contemporary records, would directly impact not only the two major coastal resorts of Santorini (Kamari and Perissa), but would also overrun the airport. In terms of the direct impact on the islands of Santorini, none of the historical eruptions within the confines of the present-day caldera has approached the scale of the Columbos Reef event and this, along with its continuing (but gentle) seismic unrest (Dimitriadis et al. 2005), makes it the focus for concern over the possibility of future activity.

100

Kos, Yali, and Nisyros

10 1

10 100 Elapsed time (days)

1000

Fig. 15.14. (a) The duration of eruptions on the Kameni islands since AD 1570 are apparently predictable, and linearly related to the time elapsed since the previous eruption, for inter-eruption periods of >10 years. This is consistent with eruption durations being controlled by the amount of magma that has accumulated in a shallow chamber between eruptions. (b) The heights of lava domes as a function of time elapsed during an eruption are shown for the 1866 and 1939 Kameni eruptions (grey circles) and compared to the domes of Mt. St Helens and St Vincent. The Kameni domes grew with approximately the same time dependence as the domes of St Helens and St Vincent, consistent with a model of steady inflation of a dome with a stiff outer shell. Since dome height is related to the duration of the eruption, the sizes of domes expected in future eruptions of the Kamenis are predictable (modified from Pyle and Elliott 2006).

violence that apparently included the ejection of large ballistic blocks, followed, over the next three days, by the emission of considerable quantities of noxious gases as well as of ash and steam. Contemporary reports suggest that the gases alone accounted for the deaths of tens of people by suffocation, as well as of countless animals. The climax of the eruption was accompanied by a great wave, usually interpreted as a tsunami, which

The three islands of Kos, Yali, and Nisyros form a cluster of recent volcanic activity in the Dodecanese, close to the Bodrum Peninsula of Turkey (Figure 15.1). Kos, the second largest of the Dodecanese islands, comprises a metamorphic basement, overlain by Mesozoic limestones and flysch. Volcanic activity in the Miocene led to the emplacement of extrusive ignimbrites and intrusive granitoids, and was followed by a period of erosion during which much of western Kos was planed to sea level (Higgins and Higgins 1996). A subsequent period of volcanism began about 3 Myr ago, and resulted in the emplacement of a couple of large rhyolite domes on the south-western Peninsula, and a major pumice dome. This phase culminated in caldera collapse and rhyolite extrusion at around 0.55 Myr ago. The most significant, and recent, activity on Kos was the eruption of the Kos Plateau Tuff at 161 kyr BP (S. R. Allen 2001; Dufek and Bergantz 2007). This eruption, which has been very precisely dated by single crystal laser fusion and Ar-Ar techniques (P. E. Smith et al. 1996), coated much of Kos, and many of the surrounding islands, with pumice fallout and pyroclastic current deposits (Figure 15.4). The discovery of Kos Plateau Tuff deposits on the Turkish mainland led to considerable interest in the question of whether pyroclastic density currents had travelled across the sea (S. R. Allen and Cas 2001). New work on the structure of the sea floor and the age of the marine sediments in

456

Clive Oppenheimer and David Pyle

the region suggests, however, that the shelf between Kos and Turkey was much shallower at the time of the eruption, and it looks less likely that the flows travelled across any substantial bodies of water (Pe-Piper et al. 2005). One curious feature of the Kos volcanic sequences is that the deposits tend to be very silica rich (rhyolitic), and the eruptions are widely spaced in time. This contrasts with the much more typical subduction volcano of Santorini, which shows both a greater diversity of melt compositions and a more systematic pattern of repeated eruption. The Kos Plateau Tuff is actually both highly crystalline and silica rich, leading to the suggestion that it may either represent a partial melt of a young granite body or have formed by remobilization of a partially consolidated magma body (Keller 1969). This process, of remobilization of previously stagnated magma bodies, appears to be common in continental margin settings, and may explain the similarities of a number of large- to very largevolume ignimbrite-forming events, including the cataclysmic eruptions of Toba (about 74 kyr ago; Gardner et al. 2002; Oppenheimer 2002) and the Fish Canyon Tuff (around 25 Myr ago; Bachmann et al. 2002). As yet, though, the processes that control when remobilization may occur remain obscure, and our understanding of the potential for future volcanic activity on Kos is limited. The small, 2 km-long, volcanic island of Yali lies south of Kos in the direction of Nisyros. This enigmatic island is thought to be a portion of a partially submerged caldera, and comprises spectacular Late Pleistocene spherulitic obsidian flows to the north, and a Quaternary sequence of pumiceous rhyolites to the south. Isotopic and trace elemental studies show that the Yali rhyolites are distinct from the rhyodacites of Nisyros (Buettner et al. 2005). The >150 m-thick pumice breccia sequence exposed on the south of the island is believed to be the uplifted remnant of a tephra rampart built by a series of submarine and phreatomagmatic eruptions (S. R. Allen and McPhie 2000). Yali has certainly been the site of Holocene activity, since the youngest pumice deposits overlie a sequence of stratified paleosols, one of which contains Neolithic pottery (Keller 1980). Today, the pumice deposits are extensively mined for a variety of commercial uses, but the obsidian, by virtue of its spherulitic nature, was never of a sufficiently high quality for use in blade manufacture (Higgins and Higgins 1996). Hydrothermal activity continues offshore from Yali, as well as in other areas between Kos and Nisyros (Varnavas and Cronan 1991), and the possibil-

ity of future volcanic activity in this region cannot be ruled out. Nisyros is currently the Aegean volcano that is showing the most unrest (Gottsmann et al. 2007). It is a textbook symmetrical stratocone rising sharply out of the sea. The subaerial part forms a small island (with an area of about 42 km2 ), dominated by a central caldera. It comprises a complex of interstratified andesitic-torhyodacitic tephra deposits, probably emplaced between around 60 and 35 kyr ago, which are partially overlain by a younger sequence of large, rhyodacite domes, which now dominate the western portion of the caldera. Historic activity has been mainly phreatomagmatic, producing the extensive fine ash deposits and explosion craters of the Lakki plains that form the eastern portion of the caldera. Nisyros has been extensively studied in the past, but the absence of such excellent and continuous exposure as exists, for example, on Santorini, has made the details of the volcanic history of the island much harder to unravel. Di Paola (1974) produced the first extensive map and descriptions of the island, while the pyroclastic record and tephra dispersal were investigated by Limburg and Varekamp (1991) and Hardiman (1999). Vougioukalakis (1993) has published the most detailed map and stratigraphy to date, and the petrology and isotope geochemistry of the sequences have been extensively described, most recently by Francalanci et al. (1995) and Buettner et al. (2005). It is clear from these studies that recycling of magmatic products— whether melts, crystalline residues from previous melts, or older crustal fragments—has been an important factor in the evolution of Nisyros magmas. This, again, makes the products intrinsically interesting to volcanologists and petrologists, but provides little concrete information for those wishing to understand the future of the magmatic system. In terms of the pyroclastic history of the island, there is still a puzzle over the origin of the caldera: despite many efforts to search for a marine tephra associated with the two major Late Pleistocene silicic explosive eruptions of Nisyros, it is only recently that a candidate Nisyros ash has been recognized in distant, north Aegean, lake and marine cores (Margari et al. 2007; Aksu et al. 2008), and there is scant evidence that either eruption was of sufficiently large magnitude to account for the size of the caldera (Hardiman 1999). Perhaps, instead, the caldera grew incrementally during episodes of both explosive and effusive activity. Of all the volcanoes in the present Aegean arc, Nisyros is the only one that shows any evidence for a present-

Volcanoes

day hydrothermal system on land (Marini et al. 1993; Chiodini et al. 2002). This is a cause for considerable concern, since isotopic studies reveal clear evidence for a magmatic signature in the currently degassing fluids (Brombach et al. 2003; Shimizu et al. 2005). Indeed, the recent seismic (1995–8) and ground deformation crises on Nisyros were accompanied by major gas geochemical changes (1997–2001), which all point towards the prospect of a renewal of the phreatic explosive activity that occurred there in the late nineteenth century (Papadopoulos et al. 1998; Sachpazi et al. 2002; Caliro et al. 2005). The onset of such activity might be challenging to forecast, although the likely focus is clearly identified as the area close to the Lofos dome on the Lakki plain (Caliro et al. 2005). The impact on the island and its tourist economy could be substantial.

Volcanic Risk Management in the Mediterranean One of the main aims of research undertaken to assess and quantify volcanic risks is of course to implement the findings in management practice and policy. Probabilistic risk assessment for current and potential future volcanic hazards is a relatively new endeavour but it is increasingly being used to inform risk management and decision-making. Such assessment makes use of information from a variety of sources, including studies of the rock record associated with past volcanism, data from monitoring networks, theoretical models, and information on the distribution and vulnerability of people or property and assets at risk. Only recently have rigorous, evidence-based approaches emerged in volcanic risk assessment (Aspinall et al. 2003; Baxter et al. 2008a ).

Characterizing, Monitoring, and Modelling Volcanic Hazards One of the first steps in volcanic hazard assessment is characterization of the past history of activity at the volcano in question, in order to build up frequencymagnitude curves for various phenomena, and thereby a statistical basis for long-range forecasting. For instance, a number of studies have attempted to characterize the volcanic hazard presented by the Campi Flegrei, including papers by Lirer et al. (2001b) and Alberico et al. (2002). One of the problems of understanding hazard in such systems where many vents are found is identifying the most likely future vent locations (e.g., see

457

Alberico et al. 2008). Based on the frequency-magnitude relationships for past eruptions, Orsi et al. (2004) suggested that the most probable ‘maximum expected event’ is a medium-magnitude explosive eruption, most likely to be focused on the north-eastern or western sectors of the caldera. It will probably alternate between magmatic and phreatomagmatic phases with the generation of tephra fallout and pyroclastic density currents. They published hazard maps indicating vent location probability, thickness of tephra fall deposits, and areas liable to inundation by pyroclastic density currents. More recently, Mastrolorenzo et al. (2008) produced probabilistic models of tephra fall for a range of Campi Flegrei eruption scenarios and showed that the hazard to buildings is high not only within the caldera but also in Naples due to the prevailing westerly winds. Similarly, Lirer et al. (2001a ) have generated longrange hazard maps for Somma-Vesuvio based on distributions of pyroclastic deposits from the main explosive events that occurred over the last 8 kyr. For tephra fallout hazard, they considered loads exceeding 300 kg m−2 as destructive, and used the relationship between load and frequency to assess relative hazards. They concluded that a total area of around 1500 km2 —which is home to nearly 2 million people—could be affected to varying degrees by the immediate impacts of a future eruption. While such approaches may be statistically robust, they need to be evaluated in the context of improved understanding of the evolution of the magmatic system that feeds the volcano (e.g. Scaillet et al. 2008). For medium- to short-range forecasting of volcanic activity, reliance shifts from the rock record to instrumental monitoring of earthquakes, ground displacements, gas emissions, and so on. Italy may justly lay claim to the establishment of the world’s first volcano observatory, which was constructed on the slopes of Vesuvio in 1841. Today, the modern Osservatorio Vesuviano (Figure 15.15a) is operated by the Istituto Nazionale di Geofisica e Vulcanologia (INGV; ). Other INGV offices include the former Istituto Internazionale per la Vulcanologia founded in Catania in 1969 and now known as INGV ‘Catania section’ (), and a centre for geochemisty based in the University of Palermo (; websites accessed 17 November 2008). In addition to fulfilling routine monitoring responsibilities, these centres and associated university-based groups are engaged in some of the most innovative volcanological research being undertaken today (Figure 15.15b). Many of the capabilities for

458

Clive Oppenheimer and David Pyle

Etna alone are highlighted in a recent publication in the American Geophysical Union’s Geophysical Monograph series (Bonaccorso et al. 2004). Routine monitoring approaches include: r Seismology, which offers an important basis for

r

r

r r

r

r

Fig. 15.15. Volcano monitoring and crisis response. (a) Operations room at the Osservatorio Vesuviano. A bank of monitors displays real-time seismological and other data streamed in from field instruments. (b) Prototype diode laser-based spectrometer for field measurement of volcanic gases being tested on the high temperature fumaroles of Gran Cratere on Vulcano. The optics for the instrument are placed directly over the hot vents with spectra recorded on a laptop in the foreground. This innovative project was one of several exploring new techniques for volcano monitoring funded by the Italian Gruppo Nazionale per la Vulcanologia (GNV). (c) Earth barriers were rapidly constructed in response to the 2001 eruption of Mt. Etna, here to deflect lava flows away from the tourist village (seen from the air in Figure 15.2b) (photos: Clive Oppenheimer).

early detection of volcanic unrest, for location of magma bodies, and for recognizing patterns related to changes in eruptive style (for example, see the work on Stromboli by Jaquet and Carniel 2003, and Langer and Falsaperla 2003). Alaprone et al. (2007) report the success of an automated alert system for tephra fallout based on seismological analysis. Geodetic survey to detect and quantify ground deformation related to magma migration below the surface. For example, both ground-based interferometric spaceborne radar methods are applied at Campi Flegrei (e.g. Lanari et al. 2004). Fluid geochemistry, which can provide evidence for influence of magma degassing in surface volcanic phenomena (e.g. Burton et al. 2003). Thermal surveys using infrared imaging devices (e.g. Calvari et al. 2005). Measurement of microgravity changes related to magmatic and hydrothermal processes (e.g. Gottsmann et al. 2003). Petrological measurements, including characterization of lava geochemistry to infer magma sources and processes (e.g. Corsaro and Miraglia 2005). Doppler radar observations of energetic eruption plumes (e.g. Donnadieu et al. 2005).

Measurements are applied in various modelling approaches. For example, during flank eruptions of Mt. Etna, lava flow hazard is modelled operationally with increasing sophistication (e.g. Barca et al. 2004; Damiani et al. 2006). Indeed there is a considerable body of research that addresses the modelling of various kinds of volcanic flow phenomena (see e.g. Sheridan 2005). Research has also been carried out on the particular behaviour of pyroclastic density currents in both the ancient (Gurioli et al. 2005) and modern (Spence et al. 2004) built environments around Vesuvius. Geographical Information Systems are also finding an increasing role in risk management (e.g. Pareschi et al. 2000a ; Renschler 2005). The specific issues of aviation risk arising from volcanic activity are addressed by the International Airways Volcano Watch (IAVW), which was established by the International Civil Aviation Organization (ICAO) in the late 1980s. The IAVW consists of

Volcanoes

several regional Volcanic Ash Advisory Centres (VAACs), which are responsible for monitoring ash clouds and delivery of warnings to the aviation community. The Mediterranean region is covered by a VAAC in Grenoble, France.

Reducing Risk Science-based evaluation of risks alone is insufficient if the end goal is to protect vulnerable populations. It is now widely acknowledged that systematic development and application of policies, strategies, and practices is essential, and that social and economic policies designed to reduce vulnerability are crucial (see Francis and Oppenheimer 2004; Chapter 18). Communities, and the officials responsible for their protection, need to appreciate not only the nature of risks, but also the uncertainties in the science underpinning the forecasting of volcanic activity and the effectiveness of actions that may be taken to prevent or mitigate the risks. Practicalities of risk management plans also need to be thought through in advance. For example, for successful evacuation, plans need to encompass communications, transport, lodging, medical care, and protection of assets. As a consequence, land use planning, public awareness programmes, education and training are increasingly being undertaken as part of volcanic risk management, for example in the communities threatened by Vesuvius. An acute problem in Italy that has contributed to the exposure of large numbers of people to volcanic hazards is the widespread illegal building known as abusivismo, which has plagued the country for half a century. This phenomenon is strongly manifested on the flanks of Vesuvius, where it has contributed very substantially to the risk from future volcanic activity, as at Campi Flegrei. The establishment in 1995 of the Vesuvius National Park, within a tight legal framework, at last seems to have limited further urban sprawl on the volcano. More recently, dramatic measures have been adopted by the Campania Regional Government aimed at achieving a gradual population decrease in the highest risk area around Somma-Vesuvio. These include cash incentives for families to relocate away from their homes in the zone considered most at risk. It is hoped that up to 100,000 people can be enticed out of this danger zone within fifteen years. In the comuni threatened by Vesuvius, public awareness is also being raised through exhibitions, and via school education. The latter is seen as a particularly effective means of outreach. A relatively small number of teachers introducing information on hazards and

459

risks into their curricula can collectively reach a very large number of pupils, who in turn will discuss what they have done at school with their families, in theory spreading positive messages about risk reduction. In the Neapolitan area the public awareness programme could do much to prepare the population in advance of any future activity of Vesuvius. An important component of risk management is the development of emergency plans to be implemented in the event of volcanic crisis. In Italy, the Dipartimento della Protezione Civile has drawn up plans for both Somma–Vesuvio and Campi Flegrei. Recent efforts have focused on evaluating and understanding community preparedness, risk perception, and vulnerability in districts threatened by Vesuvius, including surveys of awareness of the emergency plan developed for the area (e.g. Barberi et al. 2008; Carlino et al. 2008; Solana et al. 2008). While these and other studies have found evidence for a degree of hazard awareness and preparedness in the at-risk population, they have also indicated substantial scope for improvement. The Aegean has experienced far less volcanic activity than Italy in the historic period and it is not surprising therefore that volcanology, in general, is far less prolific in Greece in contrast to work on seismicity and the earthquake hazard (Chapter 16). One consequence is that planning for future volcanic emergencies in Greece appears to be rudimentary; some shortcomings have been highlighted recently for Santorini by DomineyHowes and Minos-Minopoulos (2004). Although techniques have yet to be devised that can prevent volcanic eruptions (and are frankly hard to envisage), numerous engineering solutions are available to modify hazards. Of particular note is the construction of earth barriers for lava flow control on Etna (Barberi and Carapezza 2004). Flank eruptions of the volcano are usually accompanied by remarkable efforts by the civil protection authorities to deflect lava flows away from valuable assets such as the tourist villages on the upper part of the mountain (Figure 15.15c).

Conclusions Millions of people live at risk from volcanism in the Mediterranean area. As we have seen, the hazards presented are extremely varied, and potentially devastating at the regional scale. Reflecting, in part, these high levels of risk, some of the most sophisticated and robust techniques for mapping, monitoring, and modelling volcanoes and their associated hazards and risks have been developed through research on Mediterranean

460

Clive Oppenheimer and David Pyle

volcanoes, notably as a result of major European research projects (e.g., EXPLORIS; Baxter et al. 2008b). Indeed, Etna and Vesuvius, along with Kilauea and Mount St. Helens are among the most researched volcanoes on the planet (and in the solar system for that matter). Many methods that were merely experimental a few years ago are now routinely applied to monitoring of the Italian volcanoes. Santorini, on the other hand, provides one of the most fascinating insights into disasters and human ecology in the ancient world. In the historic period, the largest and deadliest event in the region is the 1631 eruption of Vesuvio. But the ancient and prehistoric records point to even greater and more intense volcanic activity, such as the Late Pleistocene Y5 event, whose recurrence could prove catastrophic. Aside from the threat of very large magnitude eruptions (Mason et al. 2004), given the coastal and island setting of many of the Mediterranean volcanoes, one major issue that needs to be addressed further is the extent to which future eruptions or large scale flank failure could trigger a regional tsunami (Chapter 17). At Santorini, for example, eruptions of both Colombos Reef, and the caldera-collapse accompanying the Bronze Age Minoan eruption have been strongly implicated in the formation of tsunamis (e.g. Cita and Aloisi 2000; McCoy and Heiken 2000; Pareschi et al. 2006). Although the search for related tsunami deposits on the southern shores of Santorini, and the northern coast of Crete has proved equivocal to date (e.g. compare DomineyHowes 2002 and Minoura et al. 2000), it remains unclear whether this reflects a lack of preservation, less intense tsunamis than popularly imagined, or, and perhaps most likely, uncertain criteria for distinguishing the products of a tsunami event from a storm surge or turbidity current triggered by a different process. More recently, Bruins et al. (2008) have identified tsunami deposits at Palaikastro on the eastern coast of Crete that suggested wave heights of 9 m. Whatever the case for the Minoan eruption, we need only consider the welldocumented 1883 eruption of Krakatau to appreciate the potential future scenarios.

References Aiuppa, A., Moretti, R., Federico, C., Giudice, G., Gurrieri, S., Liuzzo, M., Papale, P., Shinohara, H., and Valenza, M. (2007), Forecasting Etna eruptions by real-time observation of volcanic gas composition. Geology 35: 1115–18. Alberico, I., Lirer, L. Petrosino, P., and Scandone, R. (2002), A methodology for the evaluation of long-term volcanic risk from pyroclastic flows in Campi Flegrei (Italy). Journal of Volcanology and Geothermal Research 116: 63–78.

(2008), Volcanic hazard and risk assessment from pyroclastic flows at Ischia island (southern Italy). Journal of Volcanology and Geothermal Research 171: 118–36. Allard, P., Carbonnelle, J., Dajlevic, D., Le Bronec, J., Morel, P., Robe, M. C., Maurenas, J. M., Faivre-Pierret, R., Martin, D., Sabroux, J. C., and Zettwoog, P. (1991), Eruptive and diffuse emissions of CO2 from Mount Etna. Nature 351: 387–91. Burton, M., and Muré, F. (2005), Spectroscopic evidence for a lava fountain driven by previously accumulated magmatic gas. Nature 433: 407–10. Aksu, A. E., Jenner, G., Hiscott, R. N., and Isler, E. B. (2008), Occurrence, stratigraphy and geochemistry of Late Quaternary tephra layers in the Aegean Sea and the Marmara Sea. Marine Geology 252: 174–92. Allen, A. G., Oppenheimer, C., Ferm, M., Baxter, P. J., Horrocks, L. A., Galle, B., McGonigle, A. J. S., and Duffell, H. J. (2002), Primary sulfate aerosol and associated emissions from Masaya Volcano, Nicaragua. Journal of Geophysical Research 107 (D23): 4682, doi: 10.1029/2002JD002120. Mather, T. A., McGonigle, A. J. S., Aiuppa, A., Delmelle, P., Davison, B., Bobrowski, N., Oppenheimer, C., Pyle, D. M., and Inguaggiato, S. (2006), Sources, size distribution and downwind grounding of aerosols from Mt. Etna. Journal of Geophysical Research 111: D10302, doi: 10.1029/2005JD006015. Allen, S. R. (2001), Reconstruction of a major caldera-forming eruption from pyroclastic deposit characteristics: Kos Plateau Tuff, eastern Aegean Sea. Journal of Volcanology and Geothermal Research 105: 141–62. and Cas, R. A. F. (2001), Transport of pyroclastic flows across the sea during the explosive, rhyolitic eruption of the Kos Plateau Tuff, Greece. Bulletin of Volcanology 62: 441–56. and McPhie, J. (2000), Water-settling and resedimentation of submarine rhyolitic pumice at Yali, eastern Aegean, Greece, Journal of Volcanology and Geothermal Research 95: 285–307. Alparone, S., Andronico, D., Sgroi, T., Ferrari, F., Lodato, L., and D. Reitano (2007), Alert system to mitigate tephra fallout hazards at Mt. Etna Volcano, Italy. Natural Hazards 43: 333–50. Amato, A., Chiarabba, C., Cocco, M., Dibona, M., and Selvaggi, G. (1994), The 1989–1990 seismic swarm in the Alban Hills volcanic area, central Italy. Journal of Volcanology and Geothermal Research 61: 225–37. Anastasakis, G. and Piper, D. J. W. (2005), Late Neogene evolution of the western South Aegean volcanic arc: sedimentary imprint of volcanicity around Milos. Marine Geology 215: 135– 58. Annunziatellis, A., Ciotoli, G., Lombardi, S., and Nolasco, F. (2003), Short- and long-term gas hazard: The release of toxic gases in the Alban Hills volcanic area (Central Italy). Journal of Geochemical Exploration 77: 93–108. Antonopoulos, J. (1992), The great Minoan eruption of Thera volcano and the ensuing tsunami in the Greek archipelago. Natural Hazards 5: 153–68. Anzidei, M., Carapezza, M. L., Esposito, A., Giordano, G., Lelli, M., and Tarchini, L. (2008), The Albano Maar Lake high resolution bathymetry and dissolved CO2 budget (Colli Albani volcano, Italy): Constraints to hazard evaluation. Journal of Volcanology and Geothermal Research 171: 258–68. Apuani, T., Corazzato, C., Cancelli, A., and Tibaldi, A. (2005), Stability of a collapsing volcano (Stromboli, Italy): Limit

Volcanoes equilibrium analysis and numerical modelling. Journal of Volcanology and Geothermal Research 144: 191–210. Arrighi, S., Tanguy, J. C., and Rosi, M. (2006), Eruptions of the last 2200 years at Vulcano and Vulcanello (Aeolian Islands, Italy) dated by high-accuracy archeomagnetism. Physics of the Earth and Planetary Interiors 159: 225–33. Aspinall, W. P., Woo, G., Voight, B., and Baxter, P. J. (2003), Evidence-based volcanology: Application to eruption crises. Journal of Volcanology and Geothermal Research 128: 273–85. Avino, R., Capaldi, G., and Pece, R. (1999), Radon in active volcanic areas of southern Italy. Nuovo Cimento Della Societa Italiana Di Fisica C—Geophysics and Space Physics 22: 379–85. Bachmann, O., Dungan, M. A., and Lipman, P. W. (2002), The Fish Canyon magma body, San Juan volcanic field, Colorado: rejuvenation and eruption of an upper-crustal batholith. Journal of Petrology 43: 1469–503. Baillie, M. G. L. and Munro, M. A. R. (1988), Irish tree rings, Santorini and volcanic dust veils. Nature 332: 344–6. Barberi, F. and Carapezza, M. L. (2004), The control of lava flows at Mt. Etna, in Bonaccorso et al. (2004), 357–69. Macedonio, G., Pareschi, M. T. and Santacroce, R. (1990) Mapping the tephra fallout risk: an example from Vesuvius, Italy. Nature 344: 142–4. Rosi, M., and Sodi, A. (1993), Volcanic hazard assessment at Stromboli based on review of historical data. Acta Vulcanologica 3: 173–87. Davis, M. S., Isaia, R., Nave, R., and Ricci, T. (2008), Volcanic risk perception in the Vesuvius population. Journal of Volcanology and Geothermal Research 172: 244–58. Barca, D., Crisci, G. M., Rongo, R., Di Gregorio, S., and Spataro, W. (2004), Application of cellular automata model SCIARA to the 2001 Mount Etna crisis, in Bonarccurso et al. (2004), 343–56. Baubron, J. C., Allard, P., and Toutain, J. P. (1990), Diffuse volcanic emissions of carbon dioxide from Vulcano Island, Italy. Nature 344: 51–3. Baxter, P. J., Aspinall, W. P., Neri, A., Zuccaro, G., Spence, R. J.S., Cioni, R., and Woo, G. (2008a ), Emergency planning and mitigation at Vesuvius: a new evidence-based approach. Journal of Volcanology and Geothermal Research. 178: 454–73. Neri, A., and Blong, R. (2008b), Evaluating explosive eruption risk at European volcanoes. Journal of Volcanology and Geothermal Research 178(3): v–ix. Beaubien, S. E., Ciotoli, G., and Lucchese, R. (2003), Carbon dioxide and radon gas hazard in the Alban Hills area (central Italy). Journal of Volcanology and Geothermal Research 123: 63–80. Beccaluva, L., Gabbianelli, G., Lucchini, F., Rossi, P. L., and Savelli, C. (1985), Petrology and K/Ar ages of volcanic dredged from the Eolian seamounts: implications for geodynamic evolution of the southern Tyrrhenian basin. Earth and Planetary Science Letters 74: 187–208. Bellomo, S., D’Alessandro, W., and Longo, M. (2003), Volcanogenic fluorine in rainwater around active degassing volcanoes: Mt. Etna and Stromboli Island, Italy. Science of the Total Environment 301: 175–85. Bescoby, D., Barclay, J., and Andrews, J. (2008), Saints and sinners: a tephrochronology for Late Antique landscape change in Epirus from the eruptive history of Lipari, Aeolian Islands. Journal of Archaeological Science 35: 2574–9. Biggeri, A., Pasetto, R., Belli, S., Bruno, C., Di Maria, G., Mastrantonio, M., Trinca, S., Uccelli, R., and Comba, P. (2004), Mortality from chronic obstructive pulmonary disease and pleural mesothelioma in an area contaminated by natural fiber

461

(fluoro-edenite). Scandinavian Journal of Work, Environment and Health 30: 249–52. Bonaccorso A. and Mattia, M. (2000), Deflation acting on Pantelleria Island (Sicily Channel) inferred through geodetic data. Earth and Planetary Science Letters 180: 91–101. Calvari, S., Garfì, G., Lodato, L., and Patanè, D. (2003), Dynamics of the December 2002 flank failure and tsunami at Stromboli volcano inferred by volcanological and geophysical observations. Geophysical Research Letters 30/18: 1941, doi: 10.1029/2003GL017702. Coltelli, M., Del Negro, C., and Falsaperla, S. (eds.) (2004), Mt. Etna: Volcanic Laboratory, Geophysical Monograph Series 143. American Geophysical Union, Washington DC. Bond, A. and Sparks, R. S. J. (1976), The Minoan eruption of Santorini. Journal of the Geological Society of London 132: 1–16. Bowie, C. and Bowie, S. H. U. (1991), Radon and health. Lancet 337: 409–13. Branney, M. J. and Kokelaar, B. P. (2002), Pyroclastic Density Currents and the Sedimentation of Ignimbrites, Memoir 27. Geological Society, London. Brogi, A. (2008), The structure of the Monte Amiata volcanogeothermal area (Northern Apennines, Italy): NeogeneQuaternary compression versus extension. International Journal of Earth Sciences 97: 677–703. Brombach, T., Caliro, S., Chiodini, G., Fiebig, J., Hunziker, J. C., and Raco, B. (2003), Geochemical evidence for mixing of magmatic fluids with seawater, Nisyros hydrothermal system, Greece. Bulletin of Volcanology 65: 505–16. Brown, R. J., Orsi, G., and de Vita, S. (2008), New insights into Late Pleistocene explosive volcanic activity and caldera formation on Ischia (southern Italy). Bulletin of Volcanology 70: 583–603. Bruins, H. J., MacGillivray, J. A., Synolakis, C. E., Benjamini, C., Keller, J., Kisch, H. J., Kleugel, A., and van der Plicht, J. (2008), Geoarchaeological tsunami deposits at Palaikastro (Crete) and the Late Minoan IA eruption of Santorini. Journal of Archaeological Science 35: 191–212. Buettner, A., Kleinhanns, I. C., Rufer, D., Hunziker, J. C., and Villa, I. M. (2005), Magma generation at the easternmost section of the Hellenic arc: Hf, Nd, Pb and Sr isotope geochemistry of Nisyros and Yali volcanoes (Greece). Lithos 83: 29–46. Burton, M., Allard, P., Murè, F., and Oppenheimer, C. (2003), FTIR remote sensing of fractional magma degassing at Mt. Etna, Sicily, in Volcanic Degassing, Special Publications 213. Geological Society, London, 281–93. Calanchi, N., Peccerillo, A., Tranne, C. A., Lucchini, F., Rossi, P. L., Kempton, P., Barbieri, M., and Wu, T. W. (2002), Petrology and geochemistry of volcanic rocks from the island of Panarea; implications for mantle evolution beneath the Aeolian island arc; southern Tyrrhenian Sea. Journal of Volcanology and Geothermal Research 115: 367–95. Caracausi, A., Chiodini, G., Ditta, M., Italiano, F., Longo, M., Minopoli, C., Nuccio, P. M., Paonita, A., and Rizzo A. (2004), Evidence of a recent input of magmatic gases into the quiescent volcanic edifice of Panarea, Aeolian Islands, Italy. Geophysical Research Letters 31: L07619, doi: 10.1029/2003GL019359. Chiodini, G., Galluzzo, D., Granieri, D., La Rocca, M., Saccorotti, G., and Ventura, G. (2005). Recent activity of Nisyros volcano (Greece) inferred from structural, geochemical and seismological data. Bulletin of Volcanology 67: 358–69. Calvari, S., Tanner, L. H., Groppelli, G., and Norini, G. (2004), Valle del Bove, eastern flank of Etna Volcano; a comprehensive

462

Clive Oppenheimer and David Pyle

model for the opening of the depression and implications for future hazards, in Bonaccurso et al. (2004), 65–75. Spampinato, L., Lodato, L., Harris, A. J. L., Patrick, M. R., Dehn, J., Burton, M. R., and Andronico, D. (2005), Chronology and complex volcanic processes during the 2002–2003 flank eruption at Stromboli volcano (Italy) reconstructed from direct observations and surveys with a handheld thermal camera. Journal of Geophysical Research-Solid Earth 110: B02201. Capaccioni, B., Tassi, F., Vaselli, O., Tedesco, D., and Poreda, R. (2007), Submarine gas burst at Panarea Island (southern Italy) on 3 November 2002: A magmatic versus hydrothermal episode. Journal of Geophysical Research 112: B05201, doi: 10.1029/2006JB004359. Capaldi, G., Civetta, L., and Gasparini, P. (1976/7), Volcanic history of the island of Ischia (south Italy). Bulletin of Volcanology 40: 11–22. Carapezza, M. L., Badalamenti, B., Cavarra, L., and Scalzo, A. (2003), Gas hazard assessment in a densely inhabited area of Colli Albani Volcano (Cava dei Selci, Roma). Journal of Volcanology and Geothermal Research 123: 81–94. Carlino, S., Somma, R., and Mayberry, G. C. (2008), Volcanic risk perception of young people in the urban areas of Vesuvius: Comparisons with other volcanic areas and implications for emergency management. Journal of Volcanology and Geothermal Research 172: 229–43. Catapano, F., Malafronte, R., Lepre, F., Cozzolino, P., Arnone, R., Lorenzo, E., Tartaglia, G., Starace, F., Magliano, L., and Maji, M. (2001), Psychological consequences of the 1998 landslide in Sarno, Italy: a community study. Acta Psychiatrica Scandinavica 104: 438–42. Chester, D. K., Duncan, A. M., Guest, J. E., and Kilburn, C. R. J. (1985), Mount Etna: The Anatomy of a Volcano. Stanford, Stanford University Press. Johnston, P. A., and Smolenaars, J. J. L. (2000), Human response to Etna Volcano during the classical period, in The Archaeology of Geological Catastrophes, Special Publications 171. Geological Society, London, 179–88. Dibben, C. J. L., and Duncan, A. M. (2002), Volcanic hazard assessment in Western Europe. Journal of Volcanology and Geothermal Research 115: 411–35. Duncan, A. M., and Dibben, C. J. L. (2008), The importance of religion in shaping volcanic risk perception in Italy, with special reference to Vesuvius and Etna. Journal of Volcanology and Geothermal Research 172: 216–28. Chiocci, F. L., Romagnoli, C., Tommasi, P., and Bosman, A. (2008), The Stromboli 2002 tsunamigenic submarine slide: Characteristics and possible failure mechanisms. Journal of Geophysical Research 113: B10102, doi: 10.1029/2007 JB005172. Chiodini, G., Brombach, T., Caliro, S., Cardellini, C., Marini, L., and Dietrich, W. (2002), Geochemical evidences of an ongoing volcanic unrest at Nisyros Island (Greece). Geophysical Research Letters 29: 1759, doi: 10.1029/2001GL014355. Avino, R., Brombach, T., Caliro, S., Cardellini, S., De Vita, S., Frondini, F., Granirei, D., Marotta, E., and Ventura, G. (2004), Fumarolic and diffuse soil degassing west of Mount Epomeo, Ischia, Italy. Journal of Volcanology and Geothermal Research 133: 291–309. Cimino, G. and Toscano, G. (1998), Dissolution of trace metals from lava ash: influence on the composition of rainwater in the Mount Etna volcanic area. Environmental Pollution 99: 389–93.

Cioni, R., Levi, S., and Sulpizio, R. (2000), Apulian Bronze Age pottery as a long-distance indicator of the Avellino Pumice eruption (Vesuvius, Italy), in The Archaeology of Geological Catastrophes, Special Publication 171. Geological Society, London, 159–77. Longo, A., Macedonio, G., Santacroce, R., Sbrana, A., Sulpizio, R., and Andronico, D. (2003), Assessing pyroclastic fall hazard through field data and numerical simulations; example from Vesuvius. Journal of Geophysical Research 108(B2): 2063, doi: 10.1029/2001JB000642. Cita, M. B. and Aloisi, G. (2000), Deep-sea tsunami deposits triggered by the explosion of Santorini (3500 y BP), eastern Mediterranean. Sedimentary Geology 135: 181–203. Civetta, L., Cornette, Y., Gillot, P. Y., and Orsi, G. (1988). The eruptive history of Pantelleria (Sicily Channel) in the last 50 ka. Bulletin of Volcanology 50: 47–57. Orsi, G., and Patella, D. (eds.) (2004), The Neapolitan Volcanoes: Vesuvius, Campi Flegrei and Ischia. Journal of Volcanology and Geothermal Research 133: 1–393. Coltelli, M., Del Carlo, P., and Vezzoli, L. (1998), Discovery of a Plinian basaltic eruption of Roman age at Etna volcano, Italy. Geology 26: 1095–8. (2000), Stratigraphic constraints for explosive activity in the past 100 ka at Etna Volcano, Italy. International Journal of Earth Sciences 89: 665–77. Corsaro, R. A. and Miraglia, L. (2005), Dynamics of 2004— 2005 Mt. Etna effusive eruption as inferred from petrologic monitoring. Geophysical Research Letters 32: L13302, doi: 10.1029/2005GL022347. Cortese, M., Frazzetta, G., and La Volpe, L. (1986), Volcanic history of Lipari (Aeolian Islands, Italy) during the last 10,000 years. Journal of Volcanology and Geothermal Research 27: 117–35. Crisci, G. M., De Rosa, R., Esperanca, S., Mazzuoli, R., and Sonnino, M. (1991), Temporal evolution of a three component system: the Island of Lipari (Aeolian Arc, southern Italy). Bulletin of Volcanology 53: 207–21. Di Gregorio, S., Rongo, R., Scarpelli, M., Spataro, W., and Calvari, S. (2003), Revisiting the 1669 Etnean eruptive crisis using a cellular automata model and implications for volcanic hazard in the Catania area. Journal of Volcanology and Geothermal Research 123: 211–30. Cronin, S. J., Neall, V. E., Lecointre, J. A., Hedley, M. J., and Loganathan, P. (2003), Environmental hazards of fluoride in volcanic ash: a case study from Ruapehu volcano, New Zealand. Journal of Volcanology and Geothermal Research 121: 271–91. Cubellis, E., Carlino, S., Iannuzzi, R., Luongo, G., and Obrizzo, F. (2004), Management of historical seismic data using GIS: The Island of Ischia (Southern Italy). Natural Hazards 33: 379–93. da Cologna, L. (1867), De l’influence des emanations volcaniques sur les etres organises, particulierement etudiée a Santorin pendant l’eruption de 1866. Adrien Delahaye, Paris. Damiani, M. L., Groppelli, G., Norini, G., Bertino, E., Gigliuto, A., and Nucita, A. A. (2006), Lava flow simulation model for the development of volcanic hazard maps for Mount Etna (Italy). Computers and Geosciences 32: 512–26. De Benedetti, A. A., Funicello, R., Giordano, G., Diano, G., Caprilli, E., and Paterne, M. (2008), Volcanology, history and myths of the Lake Albano maar (Colli Albani volcano, Italy). Journal of Volcanology and Geothermal Research 176: 387–406.

Volcanoes De Rosa, R., Guillou, H., Mazzuoli, R., and Ventura, G. (2003), New unspiked K-Ar ages of volcanic rocks of the central and western sector of the Aeolian Islands: reconstruction of the volcanic stages. Journal of Volcanology and Geothermal Research 120: 161–78. De Vivo, B., Rolandi, G., Gans, P. B., Calvert, A., Bohrson, W. A., Spera, F. J., and Belkin, H. E. (2001), New constraints on the pyroclastic eruptive history of the Campanian volcanic plain (Italy). Mineralogy and Petrology 73: 47–65. Deino, A. L., Orsi, G., de Vita, S., and Piochi, M. (2004), The age of the Neapolitan Yellow Tuff caldera-forming eruption (Campi Flegrei caldera—Italy) assessed by 40 Ar/39 Ar dating method. Journal of Volcanology and Geothermal Research 133: 157–70. Delmelle, P., Delfosse, T., and Delvaux, B. (2003), Sulfate, chloride and fluoride retention in Andosols exposed to volcanic acid emissions. Environmental Pollution 126: 445–57. Dibben, C. J. L. (2008), Leaving the city for the suburbs—the dominance of ‘ordinary’ decision making over volcanic risk perception in the production of volcanic risk on Mt Etna, Sicily. Journal of Volcanology and Geothermal Research 172: 288–99. Dimitriadis, I. M., Panagiotopoulos, D. G., Papazachos, C. B., Hatzidimitriou, P. M., Karagianni, E. E., Kane, I. (2005). Recent seismic activity (1994–2002) of the Santorini volcano using data from local seismological network, in M. Fytikas and G. E. Vougioukalakis, South Aegean Active Volcanic Arc: Present Knowledge and Future Perspectives, Developments in Volcanology 7. Elsevier, Amsterdam, 185–203. Di Paola, G. M. (1974), Volcanology and petrology of Nisyros island (Dodecanese, Greece). Bulletin of Volcanology 38: 944–87. Dimitriadis, I., Panagiotopoulos, D. G., Papazachos, C. B., Hatzidimtriou, P. M., Karagianni, E. E., and Kane, I. (2005), Recent seismic activity (1994–2002) of the Santorini volcano using data from local seismological network. Developments in Volcanology 7: 185–203. Dobran, F., Neri, A., and Todesco, M. (1994), Assessing the pyroclastic flow hazard at Vesuvius. Nature 367: 551–4. Dominey-Howes, D. (2002), Documentary and geological records of tsunamis in the Aegean Sea region of Greece and their potential value to risk assessment and disaster management. Natural Hazards 25: 195–224. and Minos-Minopoulos, D. (2004), Perceptions of hazard and risk on Santorini. Journal of Volcanology and Geothermal Research 137: 85–310. Papadopoulos, G. A., and Dawson, A. G. (2000), Geological and historical investigation of the 1650 Mt. Columbo (Thera Island) eruption and tsunami, Aegean Sea, Greece. Natural Hazards 21: 83–96. Donnadieu, F., Dubosclard, G., Cordesses, R., Druitt, T., Hervier, C., Kornprobst, J., Lénat, G.-F., Allard, P., and Coltelli, M. (2005), Remotely monitoring volcanic activity with groundbased doppler radar. Eos, Transactions of the American Geophysical Union 86/21: 201–8. Doumas, C. (1978), Thera and the Aegean World: Papers presented at the second international scientific congress, Santorini, Greece. Thera Foundation, London. Druitt, T. H., Mellors, R. A., Pyle, D. M., and Sparks, R. S. J. (1989), Explosive volcanism on Santorini, Greece. Geological Magazine 126: 95–126. Edwards, L., Mellors, R. M., Pyle, D. M., Sparks, R. S. J., Lanphere, M., Davies, M., and Barreiro, B. (1999), Santorini Volcano. Memoir 19. Geological Society, London.

463

Dufek, J. and Bergantz, G. W. (2007), Dynamics and deposits generated by the Kos Plateau Tuff eruption: Controls of basal particle loss on pyroclastic flow transport. Geochemistry Geophysics Geosystems 8: Q12007, doi: 10.1029/2007GC00 1741. Duncan, A. M., Dibben, C., Chester, D. K., and Guest, J. E. (1996), The 1928 eruption of Mount Etna Volcano, Sicily, and the destruction of the town of Mascali. Disasters 20: 1–20. Esposti Ongaro, T., Neri, A., Todesco, M., and Macedonio, G. (2002), Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modeling; II, Analysis of flow variables. Bulletin of Volcanology 64: 178–91. Fedele, F. G., Giaccio, B., Isaia, R., and Orsi, G. (2002), Ecosystem impact of the Campanian Ignimbrite eruption in Late Pleistocene Europe. Quaternary Research 57: 420–4. (2003), The Campanian Ignimbrite eruption, Heinrich event 4 and Paleolithic change in Europe, in A. Robock and C. Oppenheimer (eds.), Volcanism and the Earth’s Atmosphere. Geophysical Monograph 139: 301–28. Ferrari, L., Conticelli, S., Burlamacchi, L., and Manetti, P. (1996), Volcanological evolution of the Monte Amiata, southern Tuscany: new geological and petrochemical data. Acta Vulcanologica 8: 41–56. Festa, G., Zollo, A., Manfredi, G., Polese, M., and Cosenza, E. (2004), Simulation of earthquake ground motion and effects on engineering structures during the preeruptive phase of an active volcano. Bulletin of the Seismological Society of America 94: 2213–21. Fleurian D’Armenonville, Th. C. (1715), Nouveaux memoires des missions de la compagnie de Jesus, dans le Levant. Nicolas le Clerc, Paris. Fouqué, F. (1879). Santorin et ses éruptions. Masson, Paris. Francalanci, L., Varekamp, J. C., Vougioukalakis, G., Defant, M. J., Innocenti, F., and Manetti, P. (1995), Crystal retention, fractionation and crustal assimilation in a convecting magma chamber, Nisyros Volcano, Greece. Bulletin of Volcanology 56: 601–20. Francis, P. and Oppenheimer, C. (2004), Volcanoes. Oxford University Press, Oxford. Burton, M., and Oppenheimer, C. (1998), Remote measurements of volcanic gas compositions by solar FTIR spectroscopy. Nature 396: 567–70. Frattini, P., Crosta, G. B., Fusi, N., and Dal Negro, P. (2004), Shallow landslides in pyroclastic soils: a distributed modelling approach for hazard assessment. Engineering Geology 73: 277–95. Frazzetta, G. and La Volpe, L. (1991). Volcanic history and maximum expected eruption at ‘La Fossa di Vulcano’ (Aeolian Islands, Italy). Acta Vulcanologica 1: 107–13. Friedrich, W. L., Kromer, B., Friedrich, M., Heinemeier, J., Pfeiffer, T., and Talamo, S. (2006), Santorini eruption radiocarbon dated to 1627–1600 B.C. Science 312: 548. Funiciello, R., Giordano, G., and De Rita, D. (2003), The Albano maar lake (Colli Albani Volcano, Italy); recent volcanic activity and evidence of pre-Roman Age catastrophic lahar events (in Volcanic hazards; monitoring, prediction and mitigation). Journal of Volcanology and Geothermal Research 123: 43–61. Fytikas, M., Kolios, N., and Vougioukalakis, G. (1990), PostMinoan activity of the Santorini volcano: volcanic hazard and risk, forecasting possibilities, in D. A. Hardy, J. Keller,

464

Clive Oppenheimer and David Pyle

V. P. Galanopoulos, N. C. Flemming, and T. H. Druitt (eds.), Thera and the Aegean World III. Thera Foundation, London, ii. 183–98. Gardner, J. E., Layer, P. W., and Rutherford, M. J. (2002), Phenocrysts versus xenocrysts in the youngest Toba Tuff: Implications for the petrogenesis of 2800 km3 of magma. Geology 30: 347–50. Gottsmann, J., Berrino, G., Rymer, H., and Williams-Jones, G. (2003), Hazard assessment during caldera unrest at the Campi Flegrei, Italy; a contribution from gravityheight gradients. Earth and Planetary Science Letters 211: 295–309. Carniel, R., Coppo, N., Wooller, L., Hautmann, S., and Rymer, H. (2007), Oscillations in hydrothermal systems as a source of periodic unrest at caldera volcanoes: Multiparameter insights from Nisyros, Greece. Geophysical Research Letters 34: L07307, doi: 10.1029/2007GL029594. Grasso, M. F., Clocchiatti, R., Carrot, F., Deschamps, C., and Vurro, F. (1999). Lichens as bioindicators in volcanic areas: Mt. Etna and Vulcano Island (Italy). Environmental Geology 37: 207–17. Guest, J. E. and Murray, J. B. (1979), An analysis of hazard from Mount Etna Volcano. Journal of the Geological Society 136: 347–54. Cole, P. D., Duncan, A. M., and Chester, D. K. (2003), Volcanoes of Southern Italy. Geological Society, London. Gurioli, L., Pareschi, M. T., Zanella, E., Lanza, R., Deluca, E., and Bisson, M. (2005), Interaction of pyroclastic density currents with human settlements: Evidence from ancient Pompeii. Geology 33: 441–4. Hammer, C. U., Clausen, H. B., Friedrich, W. L., and Tauber, H. (1987), The Minoan eruption of Santorini in Greece dated to 1645 BC. Nature 328: 517–19. Hansell, A. L. and Oppenheimer, C. (2004), Health hazards from volcanic gases: a systematic literature review. Archives of Environmental Health 59/12: 628–39. Hardiman, J. C. (1999), Deep sea tephra from Nisyros Island, eastern Aegean Sea, Greece. Geological Society of London, Special Publication 161: 69–88. Hardy, D. A. (1990), Thera and the Aegean World III: Proceedings of the Third International Congress. Thera Foundation, London. Higgins, M. D. and Higgins, R. (1996), A Geological Companion to Greece and the Aegean. Duckworth, London. Hübner, A., Rahders, E., Rahner, S., Halbach, P., and Varnavas, S. P. (2004), Geochemistry of hydrothermally influenced sediments off Methana (western Hellenic volcanic arc). Chemie der Erde-Geochemistry 64: 75–94. Jaquet, O. and Carniel, R. (2003), Multivariate stochastic modelling: towards forecasts of paroxysmal phases at Stromboli. Journal of Volcanology and Geothermal Research 128: 261–71. Karner, D. B., Marra, F., and Renne, P. R. (2001), The history of the Monti Sabatini and Alban Hills volcanoes: groundwork for assessing volcanic-tectonic hazards for Rome. Journal of Volcanology and Geothermal Research 107: 185–215. Keller, J. (1969), ‘Origin of rhyolites by anatectic melting of granitic crustal rocks: the example of rhyolitic pumice from the island of Kos (Aegean Sea). Bulletin Volcanologique 33: 363–81. (1980), Prehistoric pumice tephra on Aegean islands, in C. Doumas (ed.), Thera and the Aegean World II. Thera Foundation, London, ii. 49–56.

Ryan, W. B. F., Ninkovitch, D., and Altherr, R. (1978), Explosive volcanic activity in the Mediterranean over the past 200,000 yr as recorded in deep-sea sediment. Geological Society of America Bulletin 89: 591–604. Kerle, N., van Wyk de Vries, B., and Oppenheimer, C. (2003), New insight into the factors leading to the 1998 flank collapse and lahar disaster at Casita volcano, Nicaragua. Bulletin of Volcanology 65: 331–45. Kilburn, C. and McGuire, B. (2001), Italian Volcanoes. Terra, Harpenden. Kilias, S. P., Naden, J., Cheliotis, I., Shepherd, T. J., Constandinidou, H., Crossing, J., and Simos, I. (2001), Epithermal gold mineralisation in the active Aegean volcanic arc: the Profitis Ilias deposit, Milos Island, Greece. Mineralium Deposita 36: 32–44. Klose, C. D. (2007), Health risk analysis of volcanic SO2 hazard on Vulcano Island (Italy). Natural Hazards 43: 303–17. Lanari, R., Berardino, P., Borgström, S., Del Gaudio, C., De Martino, P., Fornaro, G., Guarino, S., Ricciardi, G. P., Sansosti, E., and Lundgren, P. (2004), The use of IFSAR and classical geodetic techniques for caldera unrest episodes: application to the Campi Flegrei uplift event of 2000. Journal of Volcanology and Geothermal Research 133: 247–60. Langer, H. and Falsaperla, S. (2003), Seismic monitoring at Stromboli volcano (Italy): a case study for data reduction and parameter extraction. Journal of Volcanology and Geothermal Research 128: 233–45. Leycester, E. M. (1851), Some account of the volcanic group of Santorini or Thera, once called Callistê, or the Most Beautiful. Journal of the Royal Geographical Society of London 20: 1–38. Lima, A., Cicchella, D., and Di Francia, S. (2003), Natural contribution of harmful elements in thermal groundwaters of Ischia Island (southern Italy). Environmental Geology 43: 930–40. Limburg, E. M. and Varekamp, J. C. (1991), Young Pumice deposits on Nisyros, Greece. Bulletin of Volcanology 54: 68–77. Lirer, L., Munno, R., Postiglione, I., Vinci, A., and Vitelli, L. (1997), The A.D. 79 eruption as a future explosive scenario in the Vesuvian area: evaluation of associated risk. Bulletin of Volcanology 59: 112–24. Petrosino, P., Alberico, I., and Postiglione, I. (2001a ). Long-term volcanic hazard forecasts based on Somma-Vesuvio past eruptive activity. Bulletin of Volcanology 63: 45–60. (2001b), Hazard assessment at volcanic fields; the Campi Flegrei case history. Journal of Volcanology and Geothermal Research 112: 53–73. Loppi, S. and Bonini, I. (2000), Lichens and mosses as biomonitors of trace elements in areas with thermal springs and fumarole activity (Mt. Amiata, central Italy). Chemosphere 41: 1333–5. Loughlin, S. C., Baxter, P. J., Aspinall, W. P., Darroux, B., Harford, C. L., and Miller, A. D. (2002), Eyewitness accounts of the 25 June 1997 pyroclastic flows and surges at Soufrière Hills Volcano, Montserrat, and implications for disaster management, in T. H. Druitt and B. P Kokelaar, (eds.), The Eruption of Soufrière Hills Volcano, Montserrat, from 1995 to 1999. Memoir 21. Geological Society, London, 211–30. Lourens, L. J. (2004), Revised tuning of Ocean Drilling Program site 964 and KC01B (Mediterranean) and implications for the 18O, tephra, calcareous nannofossil, and geomagnetic reversal chronologies of the past 1.1 Myr. Paleoceanography 19: PA3010, doi: 10.1029/2003PA000997.

Volcanoes Luce, J. V. (1969), The End of Atlantis: New Light on an Old Legend. Thames & Hudson, London. McCoy, F. W. and Heiken, G. (2000), Tsunami generated by the Late Bronze Age eruption of Thera (Santorini), Greece. Pure and Applied Geophysics 157: 1227–56. Macedonio, G., Papale, P., Pareschi, M. T., Rosi, M., and Santacroce, R. (1994), A statistical approach to the assessment of volcanic hazard for air traffic; application to Vesuvius, Italy, in T. Casadevall (ed.), Volcanic Ash and Aviation Safety: Proceedings of the First International Symposium on Volcanic Ash and Aviation Safety. US Geological Survey Bulletin 2047: 245–52. Manasse, A. and Viti, C. (2007), Arsenic adsorption on nanocrystalline goethite: the natural example of bolar earths from Mt Amiata (Central Italy). Environmental Geology 52: 1365–74. Manning, S. W. (1999), A Test of Time: The Volcano of Thera and the Chronology and History of the Aegean and East Mediterranean in the Mid Second Millennium BC. Oxbow, Oxford. Ramsey, C. B., Doumas, C., Marketou, T., Cadogan, G., and Pearson, C. L. (2002), New evidence for an early date for the Aegean Late Bronze Age and Thera eruption. Antiquity 76: 733–44. Manning, S. W., Ramsey, C. B., Kutschera, W., Higham, T., Kromer, B., Steier, P., and Wild, E. M. (2006), Chronology for the Aegean Late Bronze Age 1700–1400 B.C. Science 321: 565–9. Maramai, A., Graziani, L., and Tinti, S. (2005a ). Tsunamis in the Aeolian Islands (southern Italy): a review. Marine Geology 215: 11–21. Alessio, G., Burrato, P., Colini, L., Cucci, L., Nappi, R., Nardi, A., and Vilardo, G. (2005b). Near- and far-field survey report of the 30 December 2002 Stromboli (Southern Italy) tsunami. Marine Geology 215: 93–106. Margari, V., Pyle, D. M., Bryant, C., and Gibbard, P. L. (2007), Mediterranean tephra stratigraphy revisited: results from a long terrestrial sequence on Lesvos Island, Greece. Journal of Volcanology and Geothermal Research 163: 34–54. Marinatos, S. (1939), The Volcanic Destruction of Minoan Crete. Antiquity 13: 425– 39. Marinelli, G. (1969), Some geological data on the geothermal areas of Tuscany. Bulletin of Volcanology 33: 319–33. Marini, L., Principe, C., Chiodini, G., Cioni, R., Fytikas, M., and Marinelli, G. (1993), Hydrothermal eruptions of Nisyros (Dodecanese, Greece). Past events and present hazard. Journal of Volcanology and Geothermal Research 56: 71–95. Martin, R. S., Mather, T. A., Pyle, D. M., Power, M., Allen, A. G., Aiuppa, A., Horwell, C. J., and Ward, E. P. W. (2008), Composition-resolved size distributions of volcanic aerosols in the Mt. Etna plumes. Journal of Geophysical Research 113: D17211. Martin, V. M., Holness, M. B., and Pyle, D. M. (2006), Textural analysis of magmatic enclaves from the Kameni Islands, Santorini, Greece. Journal of Volcanology and Geothermal Research 154: 89–102. Morgan, D. J., Jerram, D. A., Caddick, M. J., Prior, D. J., and Davidson, J. P. (2008), Bang! Month-scale eruption triggering at Santorini volcano. Science 321: 1178. Martini, M., Giannini, L., Prati, F., Tassi, F., Capaccioni, B., and Iozzelli, P. (1994), Chemical characters of crater lakes in the Azores and Italy: the anomaly of Lake Albano. Geochemical Journal 28: 173–84. Mason, B. G., Pyle, D. M., and Oppenheimer, C. (2004), The size and frequency of the largest explosive eruptions on Earth. Bulletin of Volcanology 66: 735–48.

465

Mastrolorenzo, G., Petrone, P. P., Pagano, M., Incoronato, A., Baxter, P. J., Canzanella, A., and Fattore, L. (2001), Herculaneum victims of Vesuvius in AD 79. Nature 410: 769–70. Palladino, D. M., Vecchio, G., and Taddeucci, J. (2002), The 472 AD Pollena eruption of Somma-Vesuvius (Italy) and its environmental impact at the end of the Roman Empire. Journal of Volcanology and Geothermal Research 113: 19–36. Pappalardo, L., Troise, C., Panizza, A., and De Natale, G. (2008), Probabilistic tephra hazard maps for the Neapolitan area: Quantitative volcanological study of Campi Flegrei eruptions. Journal of Geophysical Research 113: B07203, doi: 10.1029/2007JB004954. Mattia, M., Bonaccorso, A., and Guglielmino, F. (2007), Ground deformations in the Island of Pantelleria (Italy): Insights into the dynamic of the current intereruptive period. Journal of Geophysical Research 112: B11406, doi: 10.1029/2006 JB004781. Melekestsev, P., Kirianov, A., and Praslov, N. (1984), Catastrophic eruption in the region of the Flegrei fields (Italy) as a possible source of the volcanic ash in Upper Pleistocene deposits in the European part of the USSR. Vulkanologiia i Seismologiia 3: 35–44. (In Russian.) Mercalli, G. and Silvestri, O. (1891), Le eruzioni dell’isola di Vulcano, incominciate il 3 Agosto 1888 e terminate il 22 Marzo 1890. Relazione Scientifica. Annali dell’Ufficio Centrale di Meteorologia e Geodinamica 10/4: 1–213. Milia, A., Torrente, M. M., and Zuppetta, A. (2003), Offshore debris avalanches at Somma–Vesuvius volcano (Italy): implications for hazard evaluation. Journal of the Geological Society, London 160: 309–17. Minoura, K., Imamura, F., Kuran, U., Nakamura, T., Papadopoulos, G. A., Takahashi, T., and Yalciner, A. C. (2000), Discovery of Minoan tsunami deposits. Geology 28: 59–62. Montalto, A. (1995), Seismic assessment of phreatic-explosion hazard at ‘La Fossa’ Volcano (Island of Vulcano, Italy). Natural Hazards 11: 57–73. Murray, J. B. (2003), Seismicity and time-lagged lava output at Mount Etna: A new method of long-term forecasting at a destructive volcano. Geology 31: 443–6. Narcisi, B. and Vezzoli, L. (1999), Quaternary stratigraphy of distal tephra layers in the Mediterranean—an overview. Global and Planetary Change 21: 31–50. Nicholls, I. A. (1971), Petrology of Santorini volcanic rocks. Journal of Petrology 12: 67–119. Notcutt, G. and Davies, F. (1989), Accumulation of volcanogenic fluoride by vegetation—Mt Etna, Sicily. Journal of Volcanology and Geothermal Research 39: 329–33. Nowell, D. A. G., Jones, M. C., and Pyle, D. M. (2006), Episodic Quaternary volcanism in France and Germany. Journal of Quaternary Science 21/6: 645–75. Nunziante, L., Fraldi, M., Lirer, L., Petrosino, P., Scotellaro, S., and Cicirelli, C. (2003), Risk assessment of the impact of pyroclastic currents on the towns located around Vesuvio: a non-linear structural inverse analysis. Bulletin of Volcanology 65: 547–61. Oppenheimer, C. (2002), Limited global change due to the largest known Quaternary eruption, Toba approximate to 74 kyr BP? Quaternary Science Reviews 21: 1593–609. Orsi, G., Ruvo, L., and Scarpati, C. (1991), The recent explosive volcanism at Pantelleria. Geologische Rundschau 80: 187–200.

466

Clive Oppenheimer and David Pyle

Orsi, G., Civetta, L., Del Gaudio, C., De Vita, S., Di Vito, R. A., Isaia, R., Petrazzuoli, S. M., Ricciardi, G. P., and Ricco, C. (1999), Short-term ground deformations and seismicity in the resurgent Campi Flegrei caldera (Italy): an example of active block-resurgence in a densely populated area. Journal of Volcanology and Geothermal Research 91: 415–51. Di Vito, M. A., and Isaia, R. (2004), Volcanic hazard assessment at the restless Campi Flegrei caldera. Bulletin of Volcanology 66: 514–30. Palladino, D. M. and Simei, S. (2002), Three types of pyroclastic currents and their deposits; examples from the Vulsini volcanoes, Italy. Journal of Volcanology and Geothermal Research 116: 97–118. (2005), Eruptive dynamics and caldera collapse during the Onano eruption, Vulsini, Italy. Bulletin of Volcanology 67: 423–40. Palumbo, A. (1997), Chaos hides and generates order; an application to forecasting the next eruption of Vesuvius. Journal of Volcanology and Geothermal Research 79: 139–48. (1998), Long-term forecasting of the extreme eruptions of Etna. Journal of Volcanology and Geothermal Research 83: 167–72. Papadopoulos, G. A., Sachpazi, M., Panopoulous, G., and Stavrakakis, G. (1998), The volcanoseismic crisis of 1996– 97 in Nisyros, SE Aegean Sea, Greece. Terra Nova 10: 151–4. Pareschi, M. T., Favalli, M., Giannini, F., Sulpizio, R., Zanchetta, G., and Santacroce, R. (2000a ), May 5, 1998, debris flows in circum-Vesuvian areas (southern Italy): Insights for hazard assessment. Geology 28: 639–42. Cavarra, L., Favalli, M., Giannini, F., and Meriggi, A. (2000b), GIS and volcanic risk management. Natural Hazards 21: 361–79. Ranci, M., Valenza, M., and Graziani, G. (2001), Atmospheric dispersion of volcanic CO2 at Vulcano Island. Journal of Volcanology and Geothermal Research 108: 219–35. Santacroce, R., Sulpizio, R., and Zanchetta, G. (2002), Volcaniclastic debris flows in the Clanio Valley (Campania, Italy): insights for the assessment of hazard potential. Geomorphology 43: 219–31. Favalli, M., and Boschi, E. (2006), Impact of the Minoan tsunami of Santorini: Simulated scenarios in the eastern Mediterranean. Geophysical Research Letters 33: L18607, doi: 10.1029/2006GL027205. Patella, D. and Mauriello, P. (1999), The geophysical contribution to the safeguard of historical sites in active volcanic areas; the Vesuvius case-history. Journal of Applied Geophysics 41: 241–58. Paterne, M., Guichard, F., and Labeyrie, J. (1988), Explosive activity of the south Italian volcanoes during the past 80,000 years as determined by marine tephrochronology. Journal of Volcanology and Geothermal Research 34: 153–72. Pearce, J. and Boyle, P. (2005), Examining the relationship between lung cancer and radon in small areas across scotland. Health and Place 11: 275–82. Pearce, N. J. G., Westgate, J. A., Preece, S. J., Eastwood, W. J., and Perkins, W. T. (2004), Identification of Aniakchak (Alaska) tephra in Greenland ice core challenges the 1645 BC date for Minoan eruption of Santorini. Geochemistry Geophysics Geosystems 5: Q03005.

Pègues, Abbé (1842), Histoire et phénomènes du volcan et des îles volcaniques de Santorin: suivis d’un coup d’oeil sur l’état moral et religieux de la Grèce moderne. Imprimerie royale, Paris. Pe-Piper, G. and Piper, D. J. W. (2002), Igneous rocks of Greece: The anatomy of an orogen. Beiträge zur regionalen Geologie der Erde 30: 1–573. Perissoratis, C. (2005), Neotectonics and the Kos Plateau Tuff eruption of 161 ka, South Aegean arc. Journal of Volcanology and Geothermal Research 139: 315–38. Petrazzuoli, S. M. and Zuccaro, G. (2004), Structural resistance of reinforced concrete buildings under pyroclastic flows: a study of the Vesuvian area. Journal of Volcanology and Geothermal Research 133: 353–67. Pizzino, L., Galli, G., Mancini, C., Quattrocchi, F., and Scarlato, P. (2002), Natural gas hazard (CO2 , 222 Rn) within a quiescent volcanic region and its relations with tectonics; the case of the Ciampino-Marino area, Alban Hills Volcano, Italy. Natural Hazards 27: 257–87. Poli, S., Chiesa, S., Gillot, P.-Y., Guichard, F., and Vezzoli, L. (1989), Time dimension in the geochemical approach and hazard estimates of a volcanic area: The isle of Ischia case (Italy). Journal of Volcanology and Geothermal Research 36: 327–35. Principe, C., Tanguy, J. C., Arrighi, S., Paiotti, A., Le Goff, M., and Zoppi, U. (2004), Chronology of Vesuvius’ activity from A.D. 79 to 1631 based on archeomagnetism of lavas and historical sources. Bulletin of Volcanology 66: 703–24. Pyle, D. M. (1990), New volume estimates for the Minoan eruption, in D. Hardy, J. Keller, V. P. Galanopoulos, N. C. Flemming, and T. H. Druitt (eds.), Thera and the Aegean World III. Thera Foundation, London, ii. 113–21. and Elliott, J. R. (2006), Quantitative morphology, recent evolution and future activity of the Kameni islands volcano, Santorini, Greece. Geosphere 2: 253–68. Ricketts, G. D., Margari, V., van Andel, T. H., Sinitsyn, A. A., Praslov, N. D., and Lisitsyn, S. (2006), Wide dispersal and deposition of distal tephra during the Pleistocene ‘Campanian Ignimbrite/Y5’ eruption, Italy. Quaternary Science Reviews 25/21–2: 2713–28. Ramsey, C. B., Manning, S. W., and Galimberti, M. (2004), Dating the volcanic eruption at Thera. Radiocarbon 46: 325–44. Renschler, C. S. (2005), Scales and uncertainties in using models and GIS for volcano hazard prediction. Journal of Volcanology and Geothermal Research 139: 73–87. Rice, A. (2000), Rollover in volcanic crater lakes: a possible cause for Lake Nyos type disasters. Journal of Volcanology and Geothermal Research 97: 233–9. Rinaldi, M. and Venuti, M. C. (2003), The submarine eruption of the Bombarda volcano, Milos Island, Cyclades, Greece. Bulletin of Volcanology 65: 282–93. Rosi, M., Vezzoli, L., Aleotti, P., and De Censi, M. (1996), Interaction between caldera collapse and eruptive dynamics during the Campanian Ignimbrite eruption, Phlegraean Fields, Italy. Bulletin of Volcanology 57: 541–54. Sachpazi, M., Kontoes, C., Voulgaris, N., Laigle, M., Vougioukalakis, G., Sikioti, O., Stavrakakis, G., Baskoutas, J., Kalogeras, J., and Lepine, J. C. (2002), Seismological and SAR signature of unrest at Nisyros caldera, Greece. Journal of Volcanology and Geothermal Research 116: 19–33.

Volcanoes Salvi, F., Scandone, R., and Palma, C. (2006), Statistical analysis of the historical activity of Mount Etna, aimed at the evaluation of volcanic hazard. Journal of Volcanology and Geothermal Research 15: 159–68. Scaillet, B., Pichavant, M., and Cioni, R. (2008), Upward migration of Vesuvius magma chamber over the past 20,000 years. Nature 455: 216–19. Scandone, R., Arganese, G., and Galdi, F. (1993), The evaluation of volcanic risk in the Vesuvian area (in Mount Vesuvius). Journal of Volcanology and Geothermal Research 58: 263–71. Scarth, A. and Tanguy, J.-C. (2001), Volcanoes of Europe. Oxford University Press, Oxford. Schmid, M., Halbwachs, M., Wehrli, B., and Wuest, A. (2005), Weak mixing in lake kivu: New insights indicate increasing risk of uncontrolled gas eruption. Geochemistry Geophysics Geosystems 6: Q07009, doi: 10.1029/2004GC000892. Shackleton, N. J., Fairbanks, R. G., Chiu, T.-C., and Parrenin, F. (2004), Absolute calibration of the Greenland time scale: implications for Antarctic time scales and for ‰14 C. Quaternary Science Reviews 23: 1513–22. Sheridan, M. (ed.) (2005), Modeling and Simulation of Geophysical Mass Flows. Journal of Volcanology and Geothermal Research 3/1–2: 1–146. Shimizu, A., Sumino, H., Nagao, K., Notsu, K., and Mitropoulos, P. (2005), Variation in noble gas isotopic composition of gas samples from the Aegean arc, Greece. Journal of Volcanology and Geothermal Research 140: 321–39. Siebert, L. and Simkin, T. (2002), Volcanoes of the World: An Illustrated Catalog of Holocene Volcanoes and their Eruptions. Smithsonian Institution, Global Volcanism Program Digital Information Series, GVP-3, , accessed 30 October 2008. Sigurdsson, H., Devine, J. D., Tchoua, F. M., Presser, T. C., Pringle, M. K. W., and Evans, W. C. (1987), Origin of the lethal gas burst from Lake Monoun, Cameroun. Journal of Volcanology and Geothermal Research 31: 1–16. Simkin, T. and Siebert, L. (1994), Volcanoes of the World. Geoscience, Tucson. and Blong, R. (2001), Disasters: volcano fatalities— lessons from the historical record. Science 291: 255. Sinitsyn, A. A. (2003), A Palaeolithic ‘Pompeii’ at Kostenki, Russia. Antiquity 77: 9–14. Smith, P. A. and Cronan, D. S. (1983), The geochemistry of metalliferous sediments and waters associated with shallow submarine hydrothermal activity (Santorini, Aegean Sea). Chemical Geology 39: 241–62. Smith, P. E., York, D., Chen, Y., and Evensen, N. M. (1996), Single crystal Ar-40-Ar-39 dating of a late Quaternary paroxysm on Kos, Greece: Concordance of terrestrial and marine ages. Geophysical Research Letters 23: 3047–50. Solana, M. C., Kilburn, C. R. J., and Rolandi, G. (2008), Communicating eruption and hazard forecasts on Vesuvius, Southern Italy. Journal of Volcanology and Geothermal Research 172: 308–14. Sparks, R. S. J. and Wilson, C. J. N. (1990), The Minoan deposits: a review of their characteristics and interpretation, in D. Hardy, J. Keller, V. P. Galanopoulos, N. C. Flemming, and T. H. Druitt (eds.), Thera and the Aegean World III. Thera Foundation, London, ii. 89–99. Spence, R. J. S., Baxter, P. J., and Zuccaro, G. (2004), Building vulnerability and human casualty estimation

467

for a pyroclastic flow: A model and its application to Vesuvius. Journal of Volcanology and Geothermal Research 133: 321–43. Stewart, A. L. and McPhie, J. (2003), Internal structure and emplacement of an Upper Pliocene dacite cryptodome, Milos Island, Greece. Journal of Volcanology and Geothermal Research 124: 129–48. Stiros, S. C. and Chasapis, A. (2003), Geodetic monitoring of the Santorini (Thera) volcano. Survey Review 37: 84–8. Stothers, R. B. and Rampino, M. R. (1983), Volcanic eruptions in the Mediterranean before AD 630 from written and archaeological sources. Journal of Geophysical Research 88: 6357–71. Tanguy, J. C., Ribiere, C., Scarth, A., and Tjetjep, W. S. (1998), Victims from volcanic eruptions: a revised database. Bulletin of Volcanology 60: 137–44. Tibaldi, A. and Vezzoli, L. (2004), A new type of volcano flank failure: The resurgent caldera sector collapse, Ischia, Italy. Geophysical Research Letters 31: L14605, doi: 10.1029/2004GL020419. Tinti, S., Bortolucci, E., and Armigliato, A. (1999), Numerical simulation of the landslide-induced tsunami of 1988 on Vulcano Island, Italy. Bulletin of Volcanology 61: 121–37. Pagnoni, G., and Zaniboni, F. (2003), Tsunami generation in Stromboli Island and impact on the south-east Tyrrhenian coasts. Natural Hazards and Earth System Sciences 3: 299–309. Todesco, M., Neri, A., Esposti Ongaro, T., Papale, P., Macedonio, G., Santacroce, R., and Longo, A. (2002), Pyroclastic flow hazard assessment at Vesuvius (Italy) by using numerical modeling; I, Large-scale dynamics. Bulletin of Volcanology 64: 155–77. Ton-That, T., Singer, B., and Paterne, M. (2001), 40 Ar/39 Ar dating of latest Pleistocene (41 ka) marine tephra in the Mediterranean Sea: implications for global climate records. Earth and Planetary Science Letters 184: 645–58. Toyos, G., Oppenheimer, C., Sulpizio, R., Zanchetta, G., and Zuccaro, G. (2003), Modelling building vulnerability to debris flows in the Sarno area, Southern Italy, in D. Rickenmann and C. L. Chen (eds.), Debris-flow Hazards Mitigation: Mechanics, Prediction, and Assessment. Millpress, Rotterdam, 1209–20. van Hinsbergen, D. J. J., Snel, E., Garstman, S. A., Marunteanu, M., Langereis, C. G., Wortel, M. J. R., and Meulenkamp, J. E. (2004), Vertical motions in the Aegean volcanic arc: evidence for rapid subsidence preceding volcanic activity on Milos and Aegina. Marine Geology 209: 329–45. Varnavas, S. P. and Cronan, D. S. (1991), Hydrothermal metallogenic processes off the islands of Nisiros and Kos in the Hellenic Volcanic Arc. Marine Geology 99: 109–33. Vougioukalakis, G. (1993), Volcanic stratigraphy and evolution of Nisyros Island. Bulletin of the Geological Society of Greece 28: 239–58. Witham, C. S. (2005), Volcanic disasters and incidents: A new database. Journal of Volcanology and Geothermal Research 148: 191–233. Wulf, S., Kraml, M., Kuhn, T., Schwarz, M., Inthorn, M., Keller, J., Kuscu, I., and Halbach, P. (2002), Marine tephra from the Cape Riva eruption (22 ka) of Santorini in the Sea of Marmara. Marine Geology 183: 131–41.

468

Clive Oppenheimer and David Pyle

Zanchetta, G., Sulpizio, R., Pareschi. M. T., Leoni, F. M., And Santacroce, R. (2004), Characteristics of May 5–6, 1998 volcaniclastic debris flows in the Sarno area (Campania, southern Italy): relationships to structural damage and hazard zonation. Journal of Volcanology and Geothermal Research 133: 377–93.

Zollo, A., Maercklin, N., Vassallo, M., Dello Iacono, D., Virieux, J., and Gasparini, P. (2008), Seismic reflections reveal a massive melt layer feeding Campi Flegrei caldera. Geophysical Research Letters 35: L12306, doi: 10.1029/ 2008GL034242.

This chapter should be cited as follows Oppenheimer, C. and Pyle, D. M. (2009), Volcanoes, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 435–468.

16

Earthquakes Stathis Stiros

Introduction Earthquakes have played a major role in the evolution of the Mediterranean landscape. They are the most important geohazard in the region and huge sums are invested annually in seismic monitoring, hazard zoning, and earthquake prediction, and in the design of earthquake-resistant buildings and infrastructure. Large earthquakes of magnitude >7.0 have been recorded across the region and the archaeological record shows that earthquakes have posed a major hazard to human settlements for thousands of years (Ambraseys 1971; Shaw et al. 2008; Bottari et al. 2009; Figure 16.1 and Table 16.1). The study of Mediterranean seismicity started about 2,400 years ago when the first earthquake catalogue was compiled in ancient Greece (Papazachos and Papazachou 1997; Guidoboni et al. 1994). This key development predated, by several centuries, the construction of the first seismograph in China (Bullen and Bolt 1985). Since these early developments a great deal of research has been carried out to improve our understanding of earthquakes and associated hazards in the Mediterranean region and to provide protection from them. Earthquake resistant buildings—such as houses with timber bracing—were introduced in Asia Minor in the seventeenth century (Kirikov 1992; Simopoulos 1984; Stiros 1995) and the first strict anti-seismic construction regulations were implemented on the island of Levkas, Greece, in the nineteenth century under British Rule (Stiros 1995). The first ‘modern’, regional-scale earthquake maps and catalogues were compiled as early as the middle of the nineteenth century (Mallet 1858). Despite this progress, the death toll from Mediterranean earthquakes is still high and earthquakes in the region continue to surprise geoscientists. For example, the diffuse pattern of seismicity that is especially char-

acteristic of the eastern Mediterranean (Figure 16.2) is not easily reconciled with existing plate tectonic models, and many faults that are believed to demarcate plate boundaries (such as the Jordan Rift) are currently quiescent (Figure 16.3). Similarly, the 1995 Grevena-Kozani earthquake was a surprise for scientists, for it hit the heart of what was believed to be an aseismic region in northern Greece (Stiros 1998a). Furthermore, key aspects of the geodynamic background of the Mediterranean region remain a matter of debate. This chapter has three main aims. First, to outline the geodynamic setting for Mediterranean seismicity in the context of evolving ideas about plate tectonic theory. Second, to outline the methods that have been developed to study Mediterranean earthquakes and to describe their characteristics. Third, to assess the impacts of Mediterranean earthquakes from both a geomorphological and a geohazard management perspective.

The Geodynamic Background The advent of international global seismographic networks in the mid-twentieth century led to the discovery that most earthquakes are confined to narrow zones associated with lithosphere plate boundaries. The band of seismicity separating the Eurasian and African plates is clear in the Atlantic, but becomes less clear to the east as it traverses the Mediterranean to the east of Algeria (Figures 16.2 and 16.3). It becomes less clear in terms of both the distribution of epicentres and the focal mechanisms of earthquakes (McKenzie 1978; Kiratzi and Papazachos 1995). Plate boundaries can be identified in Turkey (e.g. the North and East Anatolian Faults), but they are not so clearly defined in areas such as Italy, Greece, and western Turkey, where the convergence velocities between Eurasia and Africa

470

Stathis Stiros

(b)

(a)

Fig. 16.1. (a) Ruins of buildings in the village of Verneuges, Provence, that were destroyed by the 1909 Lambesc earthquake (see Baroux et al. 2003). (b) A tombstone from the ancient Greek town of Nikomedia, in the epicentral area of the 1999 North Anatolian Fault earthquakes, Turkey, commemorating the death of two young boys and their teacher after the seismic collapse of their house in AD 120. This tombstone is in the Louvre Museum in Paris (Inscription code CIG–3293) (after Robert 1978).

TABLE 16.1. Some major or otherwise noteworthy earthquakes in the Mediterranean Date

Site

2003

Algerian coast

1999

Athens, Greece

1999

Kocaeli and Duzce earthquakes of the North Anatolian Fault, Turkey

1995

Kozani-Grevena, northern Greece

1981

Corinth, Greece

1980

Al Asnam, Algeria

1980

Irpinia (Campania-Basilicata), Italy Amorgos Island, Central Aegean

1956

1935

Coastal Libya

Effects and important features A major (M = 6.8) earthquake along the Algerian coast which produced the first known case of coastal uplift there. It caused disruption of underwater communication cables and Internet connections across a wide area. A relatively small (M = 5.9), but highly destructive earthquake (over 140 casualties) that triggered a warming of the political climate between Turkey and Greece. The surprising effect of this moderate earthquake was the very high accelerations, possibly locally exceeding 1g. Two major (M = 7.4, 7.2) and destructive (about 20,000 people killed) strike-slip faulting earthquakes along the North Anatolian Fault. The last shocks of a twentieth-century cluster of earthquakes, testifying to progressive rupturing of this fault. They also triggered a warming of the political climate between Turkey and Greece. A surprise M = 6.5 earthquake that revealed that no aseismic zones exist within a region of distributed seismicity—only earthquakes with longer recurrence intervals. A seismic sequence associated with surface ruptures, which allowed the first modelling of normal earthquakes based on a combination of geomorphological observations of seismic and long-term uplift and subsidence, and synthetic (modelled) seismograms. A destructive reverse faulting earthquake associated with surface rupture. Uplift blocked the flow of a river and formed a temporary lake. A major (M = 6.9) destructive earthquake with a high death toll (>3000); the first earthquake in the northern Mediterranean west of the Balkans associated with surface faulting. A magnitude 7.4 earthquake associated with a 20-m high tsunami. This was the largest event in the central Aegean region during the twentieth century and occurred in a region that had formerly been considered aseismic. The death toll was 53, but it would have been much higher if it had occurred 50 years later in the modern era of densely populated coastal areas. A magnitude 7+ earthquake in the apparently aseismic Libya which caused no damage or casualties, as it hit desert area with low population density.

Reference Meghraoui et al. (2004)

Gazetas et al. (2002)

Toksoz et al. (1979); Barka (1996); Stein et al. (1997); Erdik et al. (2003) Stiros (1998a)

Jackson et al. (1982)

Meghraoui et al. (1988) Westaway and Jackson (1984); Pantosti et al. (1993) Ambraseys (1960); Papadopoulos and Pavlides (1992)

Ambraseys (1984)

Earthquakes

471

TABLE 16.1. cont. Date

Site

1927

Damia, Jordan River Valley

1926

Rhodes, Aegean Sea

1909

Provence (France)

1908

Messina, Italy

1894

Locris, central Greece

1825

Levkas Island, Ionian Sea, Greece

1688

Izmir (Smyrni) Turkey

1202

Jordan Valley, Bekaa Valley (Middle East) Antioch, Syria

AD

526

AD

365

Crete, Alexandria, and the eastern Mediterranean

c.AD 250 373

BC

427/6

464

BC

Mavra Litharia, Gulf of Corinth, Greece Helike, central Greece

Locris, central Greece

Sparta, Greece

BC

c.500

BC

Megiddo, Israel

c.550

BC

Possibly Syros Island, central Aegean Jericho

c.1020

BC

c.1500

BC

Kea Island, central Aegean Sea

Effects and important features A magnitude 6 earthquake associated with a landslide that blocked the flow of the Jordan River for several days. This provides a physical explanation for the crossing of the Jordan River by the Jews in c.1020 BC, as described in biblical texts. A magnitude 7.4+ earthquake on Rhodes that led to destruction in Crete. Its effects inspired Sir Arthur Evans to assign the destruction in the palace of Knossos he was excavating to an earthquake. An earthquake of about magnitude 6 that led to deaths and major damage in several towns, mostly because of low-strength buildings. It represents the most damaging seismic event in France for centuries, and probably the first case of an earthquake in which the ‘topography effect’ was noticed, i.e. the amplification of seismic acceleration in the vicinity of sites with high topography gradient. A devastating earthquake of magnitude 7.5 which killed over 60,000 people; about 40% of the total population of the town of Messina and caused up to 70 cm of coastal subsidence. A single span (>2 km) suspension bridge is due to be constructed in the epicentral area of this earthquake. A destructive seismic sequence which inspired the first known pseudo-scientific (unsuccessful) prediction of future earthquakes. A very destructive earthquake. In its aftermath the whole of Levkas town was reconstructed using strict regulations and brilliant anti-seismic techniques that were introduced as a building code under British rule. After this destructive earthquake, stone-built houses were replaced by timber-reinforced houses that responded well to subsequent earthquakes. A large (magnitude 7.5+), destructive earthquake felt at distances of >2,000 km away. It was associated with a tsunami (Chapter 17). Near-total destruction from seismic shocks and a conflagration with possibly 250,000 people killed. In AD 587/8 another destructive earthquake killed another perhaps 60,000. A giant earthquake that became a legend; its anniversary, ‘the Day of Horror’ was commemorated for centuries in Egypt. This earthquake produced a tsunami that destroyed Egyptian towns. It has similarities with the December 2004 Asian tsunami. This earthquake, probably a series of earthquakes of magnitude over 8.5, produced >9 m uplift in Crete and nearly total destruction of towns. The 365 earthquake belongs to a cluster of major earthquakes across the eastern Mediterranean known as the ‘Early Byzantine tectonic Paroxysm’ sensu Pirazzoli (1986). Several metres of coastal uplift destroyed the harbour of the flourishing town of Aigeira, which was subsequently all but abandoned. A rather common earthquake that, several centuries later, became a legend and its effects (destruction, tsunami, subsidence of a coastal strip of land) exaggerated and explained as punishment for a sin city. One of the very first earthquakes known to have been associated with coastal subsidence and changes in the flow of springs. The only surviving entry from the first catalogue of earthquakes in Greece more than 2,000 years ago. The first case of an earthquake that triggered a revolution: seismic destruction of Sparta triggered a revolt of slaves and subordinate nations, and the Spartan regime survived thanks to the support of the rival state of Athens, which was interested in preserving social stability. A major earthquake that marked the end of a c.5,000 year old town that had survived numerous earlier earthquakes. Memories of this event are possibly imprinted in biblical texts including the Armageddon earthquake in the Revelation of St John. The first reported ‘prediction’ of an earthquake based on non-scientific reasoning. This earthquake is claimed to be responsible for collapse of the walls of Jericho, permitting its capture by the Jewish army. An example of an earthquake with wider physical effects and historical impacts: due to seismic coastal subsidence, a spring chamber was contaminated by saline water and the site was then abandoned.

Reference Nur and Ron (1996)

Ambraseys and Adams (1998); Stiros (1995) Baroux et al. (2003); and unpublished data from the author

Valensise and Pantosti (1992)

Unpublished data from the author Papazachos and Papazachou (1997); Stiros (1995) Stiros (1995) Ambraseys and Melville (1988) Guidoboni et al. (1994)

Guidoboni et al. (1994); Pirazzoli et al. (1996); Pirazzoli (1986); Stiros (2001); Stiros and Papageorgiou (2001) Stiros (1998b) This study

Guidoboni et al. (1994) Papazachos and Papazachou (1997) Thucidides, 1.101.2; Stiros (1996)

Nur and Ron (1996)

Guidoboni et al. (1994) Nur and Ron (1996) Stiros (2005)

472

Stathis Stiros

60°N

45°N

30°N

15°N

0°

30°W

15°W

0°°

15°E

30°E

45°E

Fig. 16.2. Epicentres of shallow earthquakes in the Mediterranean reported by the USGS during 1961–83 (modified from Jackson and McKenzie (1988) with additions). Epicentres are confined to a narrow zone in the Atlantic and the Red Sea, but are distributed over wide zones especially in Greece and western Turkey, making the identification of plate boundaries difficult. Areas free of earthquake epicentres exist between regions of distributed deformation including the Adriatic and the central-southern Aegean. The arrow points to the previously assumed aseismic area of western Macedonia (Greece) which was hit by the 1995 Kozani-Grevena earthquake.

become gradually higher (DeMets et al. 1990). This has led to intensive debate as to whether simple or more complicated styles of plate boundary exist between Eurasia and Africa as the Mediterranean has progressively narrowed during the final stages of the now extinct Tethys Ocean (Chapter 1). On the basis of the distribution of the focal mechanisms of earthquakes, Anderson and Jackson (1987) have shown that the Adriatic cannot be regarded as an African promontory, as was previously believed, while more recent Global Positioning System (GPS) data provide some evidence to suggest that the boundary between the African and Eurasian plates crosses Sicily (Hollenstein et al. 2003).

To address these plate boundary problems in the Aegean and western Turkey, McKenzie (1972) introduced the hypothesis of micro-plates. More recent studies, however, favour the hypothesis of continuous, distributed deformation across broad regions (Figure 16.4; Jackson 1993b) and of incipient rifting. The latter is a tendency for the formation of major faultcontrolled grabens and these are expressed through a new generation of small east–west trending faults at a strike favourable to accommodate strain from the North Anatolian Fault to the Ionian Sea which replace the older north–west trending faults (Hatzfeld 1999). This strain pattern is indeed confirmed by strike-slip

Earthquakes

473

Epicentres 1900-1965 Shallow

8>M>7 7>M>6 6>M>5

Intermediate Epicentral regions I – XVII Centuries

250 km

Fig. 16.3. Epicentres of instrumentally recorded earthquakes (M > 5) 1900–65, and areas affected by historical earthquakes in the eastern Mediterranean (modified from Ambraseys 1971). Note the absence of twentieth-century seismicity along some of the major plate boundaries such as the Dead-Sea Fault.

earthquakes along the North Anatolian Fault and by recent GPS data showing a westward motion of the Anatolian plate (McClusky et al. 2000). The geodynamic setting is complex, however, with many unresolved issues. Three main reasons for these uncertainties can be identified: 1. The radius of the Mediterranean arcs (e.g. the Aegean arc) is too short and the length of the subducting plate slab too small (about 200 km long) relative to those of typical island arcs of the Pacific and with variable dip (Hatzfeld et al. 1993). Detailed seismic profiling has indicated a spatial correlation between zones of middle crust mobilized by flow and an upper crust thinned by oro-

genic extension and expressed near the surface as an exhumed metamorphic core complex (Sachpazi et al. 1997). 2. The focal mechanisms of earthquakes used to derive plate movements in this area are rather limited in number. The older ones in particular are not well constrained and their pattern is not clear. For example, in the central part of the Hellenic arc they show east–west instead of the expected north–south compression (Taymaz et al. 1990). Many models assume that the eastern edge of the Hellenic arc is essentially a strike-slip margin (Figure 16.4) (Jackson 1993a), but the late Holocene and longer-term uplift history of Rhodes indicates an important thrusting component, as

474

Stathis Stiros

Eurasia

NAF

F

EA

Africa

DSF

Arabia

250 km

Fig. 16.4. Plate boundaries in the eastern Mediterranean (modified from Jackson 1993a). The hashed zone marks the area of distributed deformation between the North Anatolian Fault (NAF) and the Ionian Sea. The arrows indicate dominant plate movements. The East Anatolian Fault (EAF) and the Dead Sea Fault (DSF) are also shown. The solid triangles mark the Hellenic (Aegean) arc subduction zone between the African and Eurasian plates.

seen in seismic reflection profiles (Kontogianni et al. 2002; Kiratzi and Papazachos 1995). 3. Matters become rather more complicated in some key areas where deformation appears not to be controlled by seismicity and a deficit exists between deformation derived from instrumentally recorded seismicity and deformation deduced from tectonic and geodetic models. For example, in peninsular Italy or central Greece, no such deficit exists (Hunstad and England 1999; Ambraseys and Jackson 1990). However, in the Hellenic arc, where seismicity rates are high, instrumentally recorded earthquakes account for only a small part of the tectonic strain released. This indicates either dominant aseismic deformation (Jackson 1993a; Jackson and McKenzie 1988; Jenny et al. 2004), consistent with geodetic evidence (Billiris et al. 1991; Stiros 1993), or the occurrence of strong earthquakes with long recurrence intervals (Stiros 1998a; Stiros and Drakos 2006), or perhaps even earthquake swarms (Pirazzoli 1986; Nur and Cline 2000; Stiros 2001; Ambraseys 2004). Thessaly in central Greece, for example, was practically aseismic for centuries, but was marked by seven earthquakes of magnitude > 6 between 1930 and 1980 (see Papazachos and Papazachou 1997; Caputo and Helly 2005). Similarly, the Dead Sea Fault, and its northern extension up to the East Anatolian Fault, representing a plate boundary, appears almost aseismic in the instrumental seismicity records (Figures 16.2 and 16.3). They

were, however, reactivated during a cluster of strong earthquakes (magnitude up to 7.5+) between AD 1114 and 1202 (Ambraseys and Melville 1988; Ambraseys 2004). Similarly, while the Ionian Sea and the Peloponnese currently together represent one of the most seismically active parts of the Mediterranean, the scarcity of both historical seismicity data (Guidoboni et al. 1994; Papazachos and Papazachou 1997; Albini 2004) and of coastal palaeoseismicity data (Pirazzoli et al. 1996), suggests that is was rather aseismic in antiquity (Figure 16.5).

Methods for Studying Mediterranean Earthquakes Seismological (sensu stricto) methods focus on the identification of epicentres and the depth of earthquakes based on the analysis of recordings from various monitoring stations and the compilation of fault-plane solutions using various techniques (see Jackson 2001). Instrumental seismicity, however, scarcely covers the last hundred years or so and the database has to be supplemented by other approaches. Indeed, a key feature of the Mediterranean is that the seismic history of most areas can be reconstructed for thousands of years because of the rich historical records and excellent preservation of ancient remains (Ambraseys 1971; Stiros and Jones 1996; Bottari et al. 2009). This serves to counter some of the problems discussed above that have been met in the study of Mediterranean seismicity because of the tectonic complexity of the region.

Earthquakes

(a) Seismicity 1963–1988

(b) Maximum observed intensities 1700–1981

475

(c) Intense seismicity according to Strabo (1st Century AD)

IX VIII VII

50 km

Fig. 16.5. Seismicity across the Peloponnese during three periods: (a) Instrumentally derived seismicity (modified from Jackson 1993b). (b) Maximum observed earthquake intensities 1700–1981 using the MKS scale (based on IGME 1989). (c) Seismically active areas according to the first-century AD writer Strabo. There is a poor correlation between the seismicity pattern in the three periods and this may reflect time-dependent seismicity in this region.

It is not coincidental that the subdisciplines of Historical Seismology and Archaeoseismology were developed in the Mediterranean region (Jones and Stiros 2000). Historical Seismology is based on 2,500 years of historical records that yield information on seismic events

over this period. Key earthquake parameters such as epicentre location and earthquake magnitude and intensity can be estimated (Ambraseys 1971; Guidoboni et al. 1994; Stucchi and Camassi 1996). Archaeoseismology is based on ancient remains bearing traces of ancient earthquakes (Table 16.2) which typically

TABLE 16.2. A list of indicative criteria for the identification of earthquakes from archaeological data. Since various phenomena can produce similar effects to earthquakes, their identification requires that at least one of these criteria are satisfied and preferably in more than one building or site, and also that other possible causes of destruction can be excluded 1. 2. 3. 4.

Ancient buildings, walls etc. offset by seismic surface faults. Human skeletons buried under the debris of fallen buildings. Abrupt geomorphological changes that may be associated with the destruction and/or abandonment of buildings and sites. Characteristic structural damage and failure of built structures: r Displaced drums of dry masonry columns r Opened vertical joints and horizontal displacement of walls in dry masonry walls r Diagonal cracks in rigid walls r Triangular sections missing in the corners of masonry buildings r Cracks at the base or top of masonry columns and piers r Inclined or sub-vertical cracks in the upper parts of rigid arches, vaults, and domes, or their partial collapse along these cracks r Vertically displaced keystones in dry masonry arches and vaults r Several parallel fallen columns r Several fallen columns with their drums in a domino-style (‘slices of salami’) arrangement r Built structures deformed by horizontal forces so that rectangular forms are transformed to parallelograms. 5. The destruction and rapid reconstruction of sites, with the introduction of what can be regarded as ‘anti-seismic’ building construction techniques, but with no change in their overall culture. 6. Well-dated building collapse events that correlate with known earthquakes from historical sources. 7. Damage to, or destruction of, isolated buildings or entire sites and for which an earthquake provides the only feasible explanation. Sources: Modified from Stiros (1996); Karcz and Kafri (1978).

476

(a (a)

Stathis Stiros

(b)

(c)

Fig. 16.6. Examples of earthquake damage to ancient buildings: (a) Offsets of the drums of the marble columns of the Hephaisteion (Thesion) Temple near the Acropolis in Athens, the best-preserved ancient Greek temple. These offsets, assigned to an earthquake by the early seismologist Kritikos, led Galanopoulos (1956) to suggest that Athens was not aseismic, as was hitherto believed, but in the past had been shaken by destructive earthquakes. This idea was confirmed by the 1999 Athens earthquake, which, despite its relatively small (<6) magnitude, caused considerable damage and a high (>140) death toll. Such column drum offsets are a result of the oscillation of the columns by strong horizontal seismic forces (Sinopoli 1991; Stiros 1996). (b) Columns of a Roman villa in Susita, Galilee, close to the Dead Sea Transform fault, that were probably toppled during the AD 365 earthquake (after Nur and Ron 1996). (c) The parallel walls of Triolo Temple—a Greek temple at Selinunte, Sicily—that were damaged by an earthquake in the fourth-century BC (after Bottari 2003).

occurred in the last 3,000 to 4,000 years. This is possible because the occupation of many sites was permanent and because dwellings and other infrastructure (e.g. defensive walls, water channels, etc.) were constructed from stone and mortar, and even from sculpted hard rock, limestone, and marble. Many constructions had a strict geometric plan form so that any deformation associated with earthquakes or other natural effects can be determined (Lanciani 1918; Karcz et al. 1977; Stiros 1996; Hancock and Altunel 1997; Jones and Stiros 2000; Figure 16.6). Recent work by Bottari et al. (2009) based on surveys of ancient monuments in Sicily has shown that this part of the Mediterranean was struck by at least three major earthquakes between 400 BC and AD 600. The events were dated using coins, pottery, and other artefacts. Palaeoseismological studies commonly involve the study of geomorphological and stratigraphical anomalies caused by earthquakes that were associated with surface deformation and faulting. Such studies commonly involve observations from excavated trenches and they have yielded important data on, for example, the nature and extent of seismic slip and, in contexts where the age of the features can be established, recurrence intervals for major earthquakes can be estimated (Figure 16.7) (Meghraoui et al. 1988; Pantosti et al. 1993; Michetti et al. 1996; Chatzipetros et al. 1998). Another advantage for the study of seismicity in the Mediterranean is the presence of long bedrock coasts—frequently in hard carbonate rocks—that pre-

serve signs of seismic crustal uplift (Pirazzoli et al. 1982, 1996) (Figure 16.8; Chapter 13). These rocks can also preserve data on subsidence, occasionally in association with archaeological data (Flemming 1978; Pirazzoli et al. 1982). This situation is aided by the negligible tidal range (usually <10 to 20 cm) which allows unambiguous identification of small-scale coastal changes (Pirazzoli et al. 1996; Stiros et al. 1994; Chapter 13). The analysis of geodetic data has also proved to be a useful way of improving our understanding of seismic and tectonic deformation in the Mediterranean region. This method is based on comparisons of up to 100-year-old historical geodetic, triangulation, and levelling data (Stiros 1993; Hunstad and England 1999; Ward and Valensise 1989; Stiros and Drakos 2000), and the use of more recent technologies such as GPS (Billiris et al. 1991; Basili et al. 2008), and Satellite Laser Ranging data (Le Pichon et al. 1994).

The Characteristics of Mediterranean Earthquakes Geographical Distribution In very general terms, the Mediterranean is assumed to represent a plate boundary and, with the exception of its western part, its seismicity is distributed in broad zones rather than narrow bands (Figure 16.2). Thus, the distribution of earthquakes in the Mediterranean is rather

Earthquakes

477

Marzano Mt. 1582 m

Fault scarp

Carpineta Mt. 1415 m

Sele Valley 200 m

2 1

Piano di Pecore 1200 m

Fig. 16.7. A sketch to explain faulting and river response during the 1980 earthquake (M = 6.9) in southern Italy. Seismic subsidence of 50–80 cm along the 1980 fault temporarily disrupted runoff patterns from Piano de Pecore to the Sele Valley downstream. Seismic lowering led to ponding and sediment deposition in a lacustrine setting. Sediments revealed in trenches (1 and 2) excavated along the fault scarp showed that this process had followed each of four previous earthquakes in the last 8,000 years (modified from Pantosti et al. 1993).

uneven, as instrumental data covering a period longer than fifty years reveal (Figure 16.3). Some regions seem rather aseismic or with relatively low seismicity, for instance Libya (Ambraseys 1984), Egypt (Ambraseys et al. 1994) and most of Spain (Buforn et al. 1988). The bulk of the seismicity is confined to Greece, Turkey, and Italy. Even Israel, which is known to have suffered strong earthquakes in the past (Karcz et al. 1977; Amiran et al. 1994), currently seems rather aseismic (Nur and Ron 1996). Furthermore, the distribution of seismicity over broad zones has led to hypotheses for zones of distributed deformation, (Jackson 1993b), as well as for relatively aseismic blocks within them (Figure 16.4; Makropoulos and Burton 1981). The high magnitude 1995 Grevena-Kozani earthquake, in western Macedonia, Greece, however, indicated that such aseismic blocks may not exist (Stiros 1998a), and may only reflect a time-dependent seismicity (i.e. relatively long reccurence intervals of earthquakes).

Time-dependent Seismicity In some areas of high seismicity (such as central Greece) various studies have revealed that: (1) the epicentres of earthquakes are rather uniformly distributed; (2) their recurrence intervals are relatively small; and (3) the rates of seismicity are fairly constant so that instrumental seismicity data of just a few decades may be representative of the long-term pattern (Ambraseys and Jackson 1990). However, this situation is not typical of most Mediterranean regions. For example, the 1999 earthquakes in Turkey confirmed a remarkable

Fig. 16.8. Rocks uplifted during the 1953 magnitude 7.2 earthquakes in the Ionian Islands of western Greece. The lower platform corresponds to about 0.7 m of seismic uplift in 1953, while the upper platform at the foot of the mushroom-shaped rock at the harbour of Poros corresponds to a much earlier seismic uplift event which occurred between AD 350 and 710 (see Pirazzoli et al. 1996). (Photo: Stathis Stiros).

characteristic of the North Anatolian Fault—that over a period of sixty years, successive segments of this fault ruptured, causing an unusual sequence of very strong earthquakes along this fault (Toksoz et al. 1979; Barka 1996; Stein et al. 1997; Papadimitriou et al. 2001). There is also evidence that a similar sequence of earthquakes occurred along this fault during the fourth and fifth centuries AD (Stiros 2001). This sequence correlates with what has been called the Early Byzantine Tectonic Paroxysm. This was a short period of seismicity of unprecedented scale that affected the whole of

478

Stathis Stiros North Anatolian Fault earthquake sequence Aug.1999

Nov. 1999 1992

1964

1967 1957

1953

1944

1951 1943

1912

1942

1939

100 km

Istanbul M=7.6 M=7.1

M=7.0

M=7.3

M=7.1

M=7.8 M=6.9

Karliova 1949

1966

Fig. 16.9. Progressive rupturing of the North Anatolian Fault between 1939 and 1999 (modified from Cisterans et al. 2004). Predictions by Toksoz et al. (1979), Barka (1996), and Stein et al. (1997) for possible rupturing of a segment of this fault were verified by the 1999 earthquakes. Another seismic cluster along this fault was inferred for the fourth century AD on the basis of scarce historical data (Stiros 2001).

the eastern Mediterranean (Pirazzoli 1986; Chapter 13). It does not seem to be an isolated effect since, more recently, it has been established that such seismic paroxysms also occurred in the Middle East in the twelfth century AD (Ambraseys 2004), and possibly also in older and less well-known periods including the thirteenth to twelfth centuries BC in the eastern Mediterranean (Nur and Cline 2000).

Earthquake Depth The majority of seismic shocks in the Mediterranean are of shallow (<60 km) and intermediate (60–110 km) depth. Deep seismicity is mainly observed in the southern Aegean Sea (up to 200 km) (Papazachos and Papazachou 1997) and the Tyrrhenian Sea (up to 500 km) (Anderson and Jackson 1987). This deep seismicity is usually ascribed to active subduction or to the remnants of subduction in the Tyrrhenian Sea.

Fault Dimensions and Segmentation Faults of various sizes have been observed in the Mediterranean region. Some faults, namely those reflecting plate boundaries, are thousands of kilometres long and are associated with major earthquakes. The North Anatolian Fault is marked by a cluster of twentieth-century earthquakes (Figure 16.9) and is a good example. Most faults, however, are much smaller and extend up to a few tens of kilometres in length (see Basili et al. 2008). Even the long faults are discontinuous, consisting of segments the lengths of which are controlled by the depth of the seismogenic (uppermost) layer of the crust. In central Greece and western Turkey in particular, individual segments of seismic faults do not exceed 15 km in length, and are associated with earthquakes up to magnitude 7 (Ambraseys and Jackson 1990; Koukouvelas and Doutsos 1996; Jackson 2001).

The Magnitude and Destruction Potential of Mediterranean Earthquakes Strong earthquakes have struck extensive areas of the Mediterranean resulting in extensive destruction (Table 16.1). Figure 16.10 shows the widespread impact of the major eruptions in AD 1202 and 1926. However, the magnitude of the largest events, even for those that took place in the first decades of the twentieth century (e.g. the 1908 Messina, Italy, and the 1926, Rhodes Island, Greece, earthquakes), has been a matter of debate. In the case of the 1926 Rhodes earthquake (Figure 16.10), for example, magnitude estimates ranging from 6.9 to 8.7 have been proposed (Ambraseys and

Earthquakes

479

1926

1927

1202 0

500 km

Fig. 16.10. The areas affected by the 1202 (M = 7.5+), 1926 (M = 7.5), and 1927 (M = 6.4) earthquakes in the eastern Mediterranean (modified from Ambraseys and Melville 1988; Ambraseys and Adams 1998). The circle marks the epicentre of the 1926.

Adams 1998). Nonetheless, it is possible unambiguously to document the strongest earthquake that has hit the region. Historical data reporting a universal earthquake which hit most of the Mediterranean in AD 365 (Stiros 2001; Stiros and Papageorgiou 2001; Shaw et al. 2008) and elastic dislocation analysis of seismic uplift of the coasts of Crete—which were uplifted by up to 9 m (Pirazzoli et al. 1982; Kelletat 1991)—has led to the conclusion that the AD 365 earthquake (or, more accurately, a shock within this sequence) was associated with an event of minimum seismic moment 9.72 × 1028 dyne per cm. This corresponds to an earthquake with a minimum magnitude of 8.7! (Stiros and Drakos 2006). This was, however, an exceptional event along the Hellenic arc, while earthquakes in most other parts are definitely of much smaller magnitude. In the wider Aegean region large amounts of energy can be released by seismic sequences that comprise several shocks of more or less the same magnitude. This was the case, for example, during the four main shocks with magnitudes ranging between 6.2 and 7.2 during the 1953 Ionian Sea earthquakes (Papazachos and Papazachou 1997) and with the three

earthquakes with magnitude >6.0 during the 1981 Gulf of Corinth earthquakes (Jackson et al. 1982). This process is probably due to the fact that preexisting faults are too small to absorb the energy released, so faults become segmented with strain distributed in several earthquakes (Jackson 2001). This reduces the magnitude of each individual event, but in some cases the spectral characteristics of earthquakes of small magnitude influence the destruction potential of each event. The earthquakes that hit Skopje in the former Yugoslav Republic of Macedonia in 1963 and Kalamata and Athens in Greece in 1986 and 1999, respectively, are good examples. Each earthquake was relatively small (M < 6.0) but the associated shocks displayed high spectral accelerations and building oscillation was amplified (the resonance effect) leading to their destruction (Anagnostopoulos et al. 1987). It is also possible that features of the local geology and topography led to amplification of seismic movement and otherwise unexplained destruction (Gazetas et al. 2002). Another factor that has reduced the destruction potential of many large earthquakes has been the

480

Stathis Stiros

attenuation of seismic energy over long distances from the epicentre—especially since many strong events in the eastern Mediterranean take place offshore (Ambraseys et al. 1994; Papazachos and Papazachou 1997). Typically, large distant shocks affected only tall slender structures and only towers and minarets were affected in antiquity by distant earthquakes (Ambraseys et al. 1994). Modern multistorey buildings in places such as Israel, Lebanon, Egypt, and Libya, however, are very vulnerable to distant earthquakes, which, in the past, would have had only limited impacts in these areas. This type of risk is not limited to the eastern Mediterranean. Modern high-rise buildings in southern Spain and Morocco, for example, would be vulnerable to an event such as the Lisbon earthquake of 1755 (Chester 2001).

The Impact of Earthquakes on the Physical Environment The majority of earthquakes in the Mediterranean are felt only as ground-shaking events. Indeed, this is the meaning of the ancient and modern Greek word seismos, from which ‘seismic’ and ‘seismicity’ are derived. However, several earthquakes that have taken place within the last 100 to 200 years, as well as many older ones, have been associated with significant crustal deformation. This is reflected in significant changes in the topography of their epicentral areas (McKenzie 1972; Papazachos and Papazachou 1997). The Mediterranean is a key region for this area of geoscience research and this deformation is either primary, tectonic deformation or reflects secondary effects.

Tectonic Deformation Surface tectonic crustal deformation—an expression of faulting at depth at the ground surface—has been documented in various earthquakes in the Mediterranean basin (Jackson et al. 1982; King and Vita-Finzi 1981; Erdik 2001). The deformation may be predicted by elastic theory and the amplitude of deformation depends on the geometric characteristics and depth of the seismic fault (Okada 1985). This kind of deformation can result in the following: 1. Surface faulting producing vertical and/or horizontal offsets of the ground surface along the fault (Figure 16.11). Vertical displacements along normal faulting were associated with the Irpinia (1980) and the Gulf of Corinth earthquakes (Westaway and Jackson 1984; Jackson et al. 1982), and

along thrust faulting associated with the 1980 Al Asnam and older earthquakes in Algeria (King and Vita-Finzi 1981; Meghraoui et al. 1988). The strike-slip faulting associated with the Jordan Rift (Marco et al. 1997) and the North Anatolian Fault (Stein et al. 1997; Erdik 2001) are some of the bestknown examples (Figure 16.11). 2. Vertical deformation, usually in the form of uplift and/or subsidence which cannot be estimated from seismic surface ruptures. Such effects can be deduced from conventional geodetic measurements. The 1915 Avezzano and the 1980 Campania-Basilicata events in Italy and the 1978 Thessaloniki event in Greece are good examples (Ward and Valensise 1989; Stiros and Drakos 2000). Other approaches have involved a combination of GPS/triangulation and INSAR data (for instance, the 1995 Kozani, Greece, earthquake; Meyer et al. 1996; Clarke et al. 1997; and the 1995 Dinar, Turkey, earthquake, Wright et al. 1999), or more usually in relation to sea level such as the 1981 Gulf of Corinth earthquakes (Hubert et al. 1996) and the 2003 Algerian earthquake (Meghraoui et al. 2004). Systematic Mediterranean-wide studies have revealed that coastal uplift driven by seismic events is widespread (Pirazzoli et al. 1996) and occasionally of high amplitude (Chapter 13). The AD 365 earthquake, in particular, was associated with up to 9 m of uplift along a >100-km-long sector of Crete and this led to the formation of coastal plains more than 500 m wide (Figure 16.12) (Pirazzoli et al. 1982; Kelletat 1991). The AD 365 event also led to a major tsunami that struck much of the eastern Mediterranean (Stiros and Drakos 2006; Shaw et al. 2008; Chapter 17). This event also forced down-cutting in the lower reaches of many of the river channels of western Crete in response to the abrupt change in base level (Figure 16.13). The most recently recorded cases of significant earthquake-related uplift in the Mediterranean region come from the Cephalonia islands, Greece, in the Ionian Sea (Figure 16.8) and the Algerian coast (Meghraoui et al. 2004). The most important example of recent coastal land subsidence is that associated with the 1981 Gulf of Corinth earthquake (Jackson et al. 1982; Aubert et al. 1996). Interestingly, seismic coastal uplift occurs where widely used, large-scale tectonic models predict either crustal extension-associated subsidence (Flemming 1978)—such as in the central part of the Aegean (Pirazzoli et al. 1996; Stiros et al. 2000), or only minor uplift. The water line of the Roman harbour of Aigeira

Earthquakes

(a)

481

(b) D

n ista

ce 2

) (km 4

Depth (km)

0 0

Chéliff River

Elevation change (m)

1 2 6 4 2 0 -6 -4

-2 Dis

(c)

0 2 tanc e (k m)

4 6

Fig. 16.11. (a) A normal fault produced by the 1954 earthquake in Thessaly, central Greece (after Papastamatiou and Mouyiaris 1984). (b) A schematic figure to illustrate reverse faulting, folding, and uplift that dammed the flow of the Cheliff River during the 1980 Al Asnam earthquake in Algeria. The vertical movements associated with this earthquake are shown on the graph below (modified from Stein and Yeats 1989). (c) Railway tracks offset by strike-slip faulting during the 1999 Izmit earthquake in Turkey (photo: A. Barka).

in the Gulf of Corinth (Greece) is presently at the height of 4 m above mean sea level due to one or more earthquakes (Stiros 1998b), while at the nearby rocky shore, marine fossils testify to a spectacular mean average uplift rate of 3 mm per year for the late Holocene (Pirazzoli et al. 2004; Chapter 13). In many cases coastal uplift is followed by further geomorphological changes that can modify the coastal topography. For example, in response to coastal uplift, the lower reaches of streams have incised their channels in order to adjust their gradients (Figure 16.7). In some cases, however, where stream power was relatively small and incision has not kept pace with uplift, dry valleys have formed in permeable lithologies as is the case along the southern coast of the Gulf

of Corinth during the Late Pleistocene (Armijo et al. 1996). At some locations in the Mediterranean, the presence of uplifted raised beach platforms of various ages indicates the persistence of vertical seismic movements and their significance in long-term landscape development. This is the case in western Crete where Holocene, Pleistocene, and Neogene surfaces are present (Pirazzoli et al. 1982; Le Pichon and Angelier 1979). In contrast, recent seismic uplift can be evident in areas where there are no signs of older events, as is the case of the 2003 coastal uplift in Algeria (Meghraoui et al. 2004). There is some evidence that the relationship between earthquakes producing uplift and longterm landscape evolution (derived from elevated marine

482

Stathis Stiros

Iraklion

C

Falasarna

R

E

T

E

Chania

Kastelli

Rethymnon Paleochora

9?

8

Chora Sfakion

7

6

5

4

3

Matala

2 1

Isobases marking the uplift of 1550 BP in metres Emerged plains more than 500 metres wide

0

0

20 km

Fig. 16.12. Contours of uplift across western Crete resulting from the AD 365 event (modified from Kelletat 1991). See Pirazzoli et al. (1982) for a comparison as these authors present slightly different values of uplift. Emergent coastal plains greater than 500 m wide associated with this earthquake are also shown.

sediments indicating long-term uplift) in the Mediterranean was understood thousands of years ago. For instance, the fifth-century BC poet Pindar described the island of Rhodes as a plot of land emerging from the sea. Rhodes provides a very clear example of how good agreement can be achieved from different sources of evidence. In this case these are: (1) Holocene uplift rates deduced from raised shorelines yielding uplift rates of up to 0.6 mm per year (Pirazzoli et al. 1989); (2) the remains of a c.2,000-year-old harbour that is now several metres above sea level, and (3) exposed marine sediments that are widespread across the island (Kontogianni et al. 2002). Interestingly, Pindar, among other ancient authors, reported that earthquakes were responsible for the opening of a narrow valley, through which water flows to the Aegean and the transformation of a large lake into the present-day Thessaly Plain of central Greece—

the floor of which is composed of lacustrine sediments (Helly 1989). In areas of normal faulting, such effects tend to lower the relief in contrast to the activity of reverse faults, which tend to produce uplift that blocks the flow of rivers and forms lakes. The 1980 earthquake in Al Asnam, Algeria, for example, was the last in a series of thrust-type earthquakes producing uplift and transient or longer-term blocking of the flow of the Chelif River and the formation of a lake (King and Vita-Finzi 1981; Meghraoui et al. 1988).

Secondary Effects In addition to tectonically controlled crustal movements, earthquakes can produce secondary geomorphological effects, such as landslides (Chapter 6), tsunamis (Chapter 17), liquefaction, and the compaction of unconsolidated

Earthquakes

483

Fig. 16.13. Response of a stream at Sougia, on the coast of south-west Crete, to a c.6 m uplift during the AD 365 earthquake. The upper line of notches to the right and the line marking the top of the dark area to the left indicate sea level prior to the earthquake and the corresponding base level of the stream. The stream bed has adjusted to the post-seismic sea level by incision. Based on the measurements of uplift, it was found that this earthquake was of magnitude >8.5 and was associated with a thrust offshore (Stiros and Drakos 2006). The parallel lines to the right mark the remains of shorelines that were produced by a series (>10) of earthquakes. These led to subsidence (with an average of about 0.3 m for each event) over a period of 2,000 years before the AD 365 earthquake (Pirazzoli et al. 1982 1996; Kelletat 1991). This sequence testifies to a very unusual earthquake cycle (Stiros 1996). (Photo: Stathis Stiros).

sediments. Such secondary effects are mainly due to the influence of seismic waves (i.e. ground shaking) and can have very significant and long-term impacts on local communities.

Landslides Numerous earthquakes are known to have been associated with landslides (Guidoboni et al. 1994; Papazachos and Papazachou 1997; Chapter 6) and these have been shown to disrupt the flow of the River Jordan, for example (Nur and Ron 1996). Earthquake-triggered landslides may increase sediment supply to rivers and cause the burial of parts of towns, as archaeological excavations in various parts of Greece, including Olympia have shown (Stiros 1996, 2001). In Tiryns (near Mycene, southern Greece), in particular, the burial of the lower town by debris seems to have followed the seismic

destruction of a dam about 2,800 years ago (Kilian 1996).

Tsunamis The magnitude 7.4 earthquake that struck Amorgos Island in the central Aegean Sea in 1956 produced a tsunami wave up to 20 m high (Ambraseys 1960). The tsunami associated with the AD 365 earthquake was responsible for the inundation of extensive areas of coastal land in Egypt and the transformation of cultivated land into swamps (Guidoboni et al. 1994; Stiros 2001; Shaw et al. 2008). This process was evident during the 2004 tsunami in south-east Asia. The Syrian coast has been particularly affected by tsunamis produced by earthquake-triggered submarine slumping (Chapter 17).

484

Stathis Stiros

Submerged Settlements in Antiquity

Fig. 16.14. An early engraving illustrating the combined effects of ground sliding, compaction, and perhaps liquefaction following the 1783 earthquake in Calabria, Italy (after Michetti 1994).

Compaction and Liquefaction of Unconsolidated Sediments Earthquake shaking can lead to the compaction of unconsolidated sediments sediments (see e.g. Papathanassiou et al. 2005) and this in turn can lead to ground subsidence. Liquefaction is a process whereby unconsolidated surficial sediments lose their cohesion and rigidity and this can have a highly destructive impact on overlying structures (Figure 16.14).

Hydrological Effects Strong ground shaking can change the flow rates from springs and initiate new springs as underground drainage networks become diverted and disconnected or blocked with debris (see Papazachos and Papazachou 1997).

Sinkholes Seismic events may lead to the weakening and collapse of bedrock ceilings above underground voids in karst terrains to form sinkholes. Large boulders in rockshelter and cave sites in the Mediterranean (e.g. Franchthi Cave in southern Greece) have been attributed to seismic shaking (Farrand 2000).

Other Secondary Effects Tectonic deformation may have a very wide number of indirect effects on geomorphological processes. For example, even a relatively small scale (c.1 m) uplift or subsidence at a coastal site may change the direction and magnitude of stream discharge and modify erosion and sedimentation dynamics in the catchment and at the coast.

In some cases there are historical reports of catastrophic earthquake-associated landscape changes and these commonly refer to ancient sites or parts of towns ‘swallowed by the waves’ (see Papazachos and Papazachou 1997; Nikonov 1996). However, many of these claims cannot be accepted as reliable. The best-known case is that of Helike, which along with the nearby town of Boura, lying on a hill, was supposed to have been submerged in the Gulf of Corinth with all its inhabitants during the 373 BC earthquake and tsunami that followed (Marinatos 1960). Such an extraordinary calamity, however, was first mentioned several centuries later in the case of Helike and became a legend. It was not mentioned by any of the contemporary historians, who only provided brief and rather vague reports of an earthquake followed by a small tsunami—events not unusual in the region. This author believes that the legend of the earthquake-related drowning of Helike and Boura, among other similar legends of destruction in the region, should be discarded for the following reasons: 1. Greek towns of this period covered extensive areas (ancient Athens with its hundred demoi, or communities, is the typical example) and no signs of the associated large-scale geomorphological changes that are needed to account for the subsidence of both Helike (located in the coastal plain) and of Boura (located on a hill) have ever been identified. 2. It is very likely that the ‘drowning of Helike’ is a myth that was first propagated (possibly following a low-magnitude earthquake), to disguise its political annihilation by the nearby rival town of Aigion which, miraculously, was not affected by the hypothetical giant earthquake and tsunami, and became the dominant town in the area (Faraklas 1998). 3. There are no natural events or combinations of events (e.g. seismic faulting, landslide, tsunami inundation, liquefaction of loose sediments) that can adequately explain the simultaneous loss of the two towns in their contrasting topographic settings. 4. There is some stratigraphic evidence for buildings and undisturbed occupation levels that predate and postdate the critical period (373 BC) in the region inland (Soter and Katsonopoulou 1999), and this argues strongly against the occurrence of a catastrophic landslide, land subsidence, or any other similar calamity.

Earthquakes

The Historical and Social Significance of Earthquakes in the Mediterranean Earthquakes have often been invoked as a convenient solution to explain discontinuities in an occupation sequence or in the cultural history of various sites and regions and this approach has been criticized by various investigators (Ambraseys 1971; Stiros 1996: Faraklas 1998). This is not to say that, in numerous cases, earthquakes did not play an important role in the history of some sites by, for example, influencing their political role, their architectural and urban (re)modelling, or even their abandonment. The 464 BC earthquake that destroyed Sparta triggered a revolt of slaves and oppressed tribes, and a war that risked destroying the Spartan state. Also, following its demise due to the AD 365 earthquake, Paphos was replaced by Salamis as the capital of Cyprus (Stiros 2001), while after destruction by an earthquake the town of Al Asnam in Algeria was re-established but with a different name. Archaeological excavations have indicated a radical remodelling of some towns in the aftermath of earthquakes (Stiros and Jones 1996). The redesign of Troy prior to the period corresponding to the Trojan War was assigned by Blegen (1963) to the effects of an earthquake. Similarly, the Roman town of Kisamos in Crete that was levelled during the AD 365 earthquake was totally reconstructed as a new, Christian town (Stiros and Papageorgiou 2001). Some modern examples include Salon de Provence in southern France, the centre of which was remodelled after the 1909 earthquake, and some modern Greek earthquakes such as the 1995 Aigion earthquake in the Gulf of Corinth (Koukouvelas and Doutsos 1996). The latter provided an opportunity for the people of this town to replace their traditional twostorey houses made of stone and unbaked bricks even though they formed part of their cultural heritage and were protected by law. There is an important lesson here because they were replaced by modern, multistorey, reinforced-concrete buildings that, in many cases, proved more vulnerable to earthquake destruction than the well-preserved traditional dwellings (Figure 16.15). Finally, there are numerous cases of sites and towns abandoned or relocated following earthquakes. Pella in Macedonia, northern Greece, the birthplace of Alexander the Great, was relocated a few hundred metres away from its original site after a destructive earthquake in around 90 BC (Stiros 1998a). Of the numer-

485

ous archaeological sites and ruins currently bearing names with the prefix ‘Palaio-’, meaning ‘old’, many reflect abandonment after a seismic disaster. Santorini (Thera) Island was largely abandoned after the 1956 earthquakes, until the tourist boom of the 1980s, while after the 1881 Chios Island (central Aegean Sea) earthquakes, the major part of the population left the island, ignoring the orders of the Turkish Government trying to avoid its abandonment. The same situation arose after the 1953 earthquakes in the Ionian Islands. In order to avoid the vacation of the devastated islands, the Greek Government confiscated the fishing boats, thereby condemning many fishermen to starvation. In fact the role of the central government and of other communities has been decisive in the post-earthquake history of towns. During the 464 BC earthquake that devastated Sparta, the Athenians sent an army to support their eternal rivals to prevent a slave revolt and to maintain the oppression of conquered nations. Several ancient towns are known to have survived seismic disasters thanks to financial assistance from the central government or other communities or colonies (Guidoboni 1989). Without such assistance, many settlements effectively disintegrated and ceased to function as organized towns. A good example is Gortyn (the former capital of Crete), Egypt, and Libya during Roman times (Di Vita 1996). In some cases, however, earthquakes were responsible for natural effects that had a catalyzing role in the inhabitation history of sites. For example, around 3,500 years BP, the prehistoric site of Kea Island in the central Aegean was effectively abandoned when a spring that provided the major source of potable water was contaminated by saline water following coastal seismic subsidence (Stiros 2005). Finally, it has been argued that the passage of the Jews to Canaan, as described by the Bible, was in fact made possible because of an earthquakeformed landslide dam across the Jordan River, similar to the dam that formed during an early twentieth century earthquake (Nur and Ron 1996).

Hazard Management: Defence Against Earthquakes In some parts of the world, in the aftermath of a destructive earthquake, many people have little choice but to remain in the affected region because of restrictions imposed by the availability of suitable land and building materials and the nature of the climate. New houses are commonly built in the same locations using the same techniques, even if they have proved susceptible

486

Stathis Stiros

(a)

(b)

Fig. 16.15. (a) A collapsed multistorey building that led to several casualties in Kalamata, southern Greece, following the 1986 earthquake. (b) During the same event many traditional buildings were badly damaged but did not collapse, and many of their occupants survived (see Anagnostopoulos et al. 1987). (Photos: Stathis Stiros).

to earthquakes in the past. The settlements that comprise vaulted, unbaked brick houses in the central, arid part of Iran are a good example (see Jackson 2001). In contrast, in most parts of the Mediterranean the availability of land, building materials, and good trading contacts with other regions has allowed people to invent, test, and adopt techniques for anti-seismic construction to reduce the vulnerability of their dwellings. The use of timber framing in houses and iron clamps in the hewn marble blocks of ancient temples are among the most effective anti-seismic protection techniques that have been employed (Kirikov 1992; Stiros 1995). In some cases, it is known that entire towns were rebuilt after earthquakes abandoning the old building styles and introducing anti-seismic construction techniques. Indeed, this was the case in late seventeenth-century Smyrni (Izmir) in western Turkey, after the earthquake of 1688 (Simopoulos 1984). Due to the high death toll in many Mediterranean countries with over 1,000 in Turkey and 19 per year in Greece, on average, throughout the twentieth century (Erdik et al. 2003; A. Anagnostopoulos 2003), seismic protection has become a government priority in most countries. Government-funded research in Greece has focused on understanding earthquakes, including both their history and geography (i.e. the compilation of earthquake catalogues based on instrumental (Macropoulos and Burton 1981), historical (Guidoboni et al. 1994; Ambraseys et al. 1994), or a variety of data (Amiran et al. 1994; Papazachos and Papazachou 1997) and the identification of potential seismic sources (for instance Stein et al. 1997; Valensise and Pantosti

2001). Investment has also been directed towards the compilation of maps of active faults (for instance in Turkey, see Saroglu et al. 1992) of observed seismic intensities (e.g. in Greece: IGME 1989) and of seismic risk, at local, national, and international levels (Erdik et al. 2003; Jimenez et al. 2003). Particular attention has been given to the introduction of regulations for earthquake-resistant buildings and for the development of emergency plans. In Greece an effective anti-seismic protection building code was first introduced on the island of Levkas in the Ionian Sea during the nineteenth century under British Rule (Stiros 1995), while more formal antiseismic building codes were introduced after the 1928 Corinth earthquakes. These were inspired partly by the Japanese codes and were adopted nationally across Greece in 1959. More comprehensive codes to deal with the reinforced concrete buildings that represent the great majority of modern construction in Greece were introduced in 1984 and 1992 (see A. Anagnostopoulos 2003). These codes comprise a map of seismic risk zonation and a list of regulations for the design of earthquakeresistant buildings. Seismic risk zonation is a primary responsibility of seismologists, geologists, and engineers. It is based on our understanding of past earthquakes and incorporates statistical estimates of future earthquakes—mainly estimates of recurrence intervals and expected maximum seismic ground acceleration. While efforts have been made to estimate the seismic risk in the Mediterranean using both time-dependent and time-independent seismicity models (Jimenez et al. 2003;

Earthquakes

Papaioannou and Papazachos 2000), it is difficult to accommodate in these models the time-dependent contrasts in seismic history as exemplified in western Crete, for example. While coastal and archaeological data provide evidence of a magnitude >8.7 earthquake in AD 365 (Pirazzoli et al. 1982; Kelletat 1991; Stiros and Papageorgiou 2001; Stiros and Drakos 2006; Shaw et al. 2008), the seismotectonic map of Greece, based on recorded seismic intensities in the last three hundred years, classifies this area as of low seismic risk (IGME 1989)! Seismic zoning maps are updated from time to time following major earthquakes (e.g Anastasiadis et al. 2001) or the acquisition of new data from palaeoseismic studies. Seismic zoning is not limited to earthquakeprone countries such as Italy and Turkey (Turkish General Directorate of Maps 1972, 1996), but has been adopted for countries such as France, which have much lower levels of seismicity (Desperoux and Godefroy 1986; Figure 16.2). Regulations for the construction of earthquake-resistant buildings (mostly relating to reinforced concrete structures) were first introduced in Greece and later by most Mediterranean countries. Recently there has been a move towards the adoption of unified codes (Eurocodes) that are similar to those in Japan and the United States. However, in most Mediterranean countries such codes are not mandatory and their use is commonly limited to new or remodelled buildings. Much of the existing infrastructure across the Mediterranean region therefore remains vulnerable to the earthquake hazard. It is worth pointing out that seismic risk analysis, including the identification of fault segments with the potential to rupture in a future earthquake—as is the case with the NAF (Toksoz et al. 1979; Barka 1996; Stein et al. 1997)—has little in common with various efforts for ‘short-term’ prediction, which, it can be argued, have largely proved unsuccessful (Mulargia and Geller 2003). Finally, although there have been concerted efforts for international collaboration—especially within the European Union— as far as earthquake research (seismic risk estimation), preparedness (earthquake-resistant infrastructure and emergency planning), and post-event recovery (crisis management and long-term recovery) are concerned (ibid.), there is still much work to do.

References Albini, P. (2004), A survey of past earthquakes in the Eastern Adriatic (14th to early 19th century). Annals of Geophysics 47: 675–704.

487

Ambraseys, N. (1960), The seismic sea-wave of July 9, 1956 in the Greek Archipelago. Journal of Geophysical Research 65: 1257–65. (1971), Value of historical records of earthquakes. Nature 232: 375–9. (1984), Material for the investigation of the seismicity of Tripolitania (Libya), in A. Brambati and D. Sleiko (eds.), The O.G.S. Silver Anniversary Volume. OGS Trieste, 143–53. (2004), The 12th century seismic paroxysm in the Middle East: a historical perspective. Annali di Geofisica 47: 733–58. and Adams, R. (1998), The Rhodes earthquake of 26 June 1926. Journal of Seismology 2: 267–92. and Jackson, J. (1990), Seismicity and associated strain of central Greece between 1890 and 1988. Geophysical Journal International 101: 663–708. and Melville, C. (1988), An analysis of the Eastern Mediterranean earthquake of 20 May 1202, in W. H. K. Lee, H. Meyers, and K. Shimazaki (eds.), Historical Seismograms and Earthquakes of the World. Academic Press, San Diego, 181–200. and Adams, R. (1994), The Seismicity of Egypt, Arabia and the Red Sea. Cambridge University Press, Cambridge. Amiran, D., Arieh, E., and Turcotte, T. (1994), Earthquakes in Israel and Adjacent Areas: Macroseismic observations since 100 BC. Israel Exploration Journal 44: 260–305. Anagnostopoulos, A. (2003), Seismic risk mitigation—Greek case study, in F. Mulargia and R. Geller (eds.), Earthquake Science and Seismic Risk Reduction. NATO Science Series IV. Earth and Environmental Sciences 32, 250–62. Anagnostopoulos, S., Rinaldis, D., Lekidis, V., and Theodulidis, N. (1987), The Kalamata, Greece, earthquake of September 13, 1986. Earthquake Spectra 3: 365–402. Anastasiadis, A., Raptakis, D., and Pitilakis, K. (2001), Thessaloniki’s detailed microzoning: Subsurface structure as basis for site response analysis. Pure and Applied Geophysics 158: 2597–633. Anderson, H. and Jackson, J. (1987), Active tectonics of the Adriatic Region. Geophysical Journal of the Royal Astronomical Society 91: 937–83. Armijo, A., Meyer, B., King, G., Rigo, A., and Papanastasiou, D. (1996). Quaternary evolution of the Corinth Rift and its implications for the Late Cenozoic evolution of the Aegean. Geophysical Journal International 126: 11–53. Barka, A. (1996), Slip distribution along the North Anatolian Fault associated with large earthquakes of the period 1939–1967. Bulletin of the Seismological Society of America 86: 1238–54. Baroux, E., Pino, N., Valensise, G., Scotti, O., and Cushing, M. (2003), Source parameters of the 11 June 1909, Lambesc (Provence, southeastern France) earthquake: A reappraisal based on macroseismic, seismological, and geodetic observations. Journal of Geophysical Research 108: 2454, doi: 10.1029/2002JB002348). Basili, R., Valensisea, G., Vannolia, P., Burratoa, P., Fracassia, U., Marianoa, S., Tibertia, M. M., and Boschia, E. (2008), The database of individual seismogenic sources (DISS), version 3: summarizing 20 years of research on Italy’s earthquake geology. Tectonophysics 453: 20–43. Billiris, H., Paradissis, D., Veis, G., England, P., Featherstone, W., Parsons, B., Cross, P., Rands, P., Rayson, M., Sellers, P., Ashkenazi, V., Davison, M., Jackson, J., and Ambraseys, N. (1991), Geodetic determination of tectonic deformation in central Greece from 1900 to 1988. Nature 350: 124–9.

488

Stathis Stiros

Blegen, C. (1963), Troy and the Trojans. Thames & Hudson, London. Bottari, C. (2003), Unknown earthquakes in Sicily and Calabria before AD 1000. Ph.D. Thesis, Universita degli Studi di Messina. Bottari, C., Stiros, S. C., and Teramo, A. (2009), Archaeological evidence for destructive earthquakes in Sicily between 400 BC and AD 600. Geoarchaeology 24: 147–75. Buforn, E., Udias, A., and Mezcua, J. (1988), Seismicity and focal mechanisms in South Spain—Bulletin of the Seismological Society of America 78: 2008–24. Bullen, K. and Bolt, B. (1985), An Introduction to the Theory of Seismology. 4th edn. Cambridge University Press, Cambridge. Caputo, R. and Helly, B. (2005), Archaeological evidences of past earthquakes: a contribution of the SHA of Thessaly, Greece. Journal of Earthquake Engineering 9: 199–222. Chatzipetros, A., Pavlides, S., and Mountrakis, S. (1998), Understanding the 13 May 1995 Western Macedonia Earthquake: A paleoseismological approach. Journal of Geodynamics 26: 327–39. Chester, D. (2001), The 1755 Lisbon earthquake. Progress in Physical Geography 25: 363–83. Clarke, P., Paradissis, D., Briole, P., England, P., Parsons, B., Billiris, H., Veis, G., and Ruegg, J. (1997), Geodetic investigation of the 13 May 1995 Kozani-Grevena (Greece) earthquake. Geophysical Research Letters 24: 707–10. DeMets, C., Gordon, R., Argus, D., and Stein, S. (1990), Current plate motions. Geophysical Journal International 101: 425–78. Desperoux, J. and Godefroy, P. (1986), Nouveau zonage sismique de France. La documentation Française, Paris. Di Vita, A. (1996), Earthquakes and civil life at Gortyn (Crete) in the period between Justinian and Constant II (6–7th century AD), in S. C. Stiros and R. E. Jones (eds.), Archaeoseismology, Fitch Laboratory Occasional Paper 7. Oxbow, Oxford, 45–50. Erdik, M. (2001), Report on the Kocaeli and Duzce (Turkey) earthquakes. <www.iiasa.ac.at/Research/RMS/ july2000/Papers/erdik.pdf>, accessed 4 November 2008. Durukal, E., Siyahi, B., Fahjan, J., Sesetyan, K., Demircioglu, M., and Akman, H. (2003), Seismic risk mitigation— Istanbul case study, in F. Mulargia and R. Geller (eds.), Earthquake Science and Seismic Risk Reduction. NATO Science Series, IV. Earth and Environmental Sciences 32: 262–81. Faraklas, N. (1998), Helike: Its natural and political submergence [in Greek], in Proceedings: Helike II, Ancient Helike and Aigialeia. Athens, 209–34. Farrand, W. (2000), Depositional History of Franchthi Cave. Sediments, Stratigraphy and Chronology. Excavations at Franchthi Cave, Greece 12. Indiana University Press, Bloomington. Flemming, N. (1978), Holocene eustatic changes and coastal tectonics in the Northeast Mediterranean: implications for models of crustal consumption. Philosophical Transactions of the Royal Society of London, Series A 289: 405–58. Galanopoulos, A. (1956), The seismic risk at Athens [in Greek]. Praktika Akadimias Athinon 31: 464–72. Gazetas, G., Kallou, P., and Psarropoulos, P. (2002), Topography and soil effects in the Ms 5.9 Parnitha (Athens) earthquake: The case of Adames. Natural Hazards 27: 133–69. Guidoboni, E. (ed.) (1989), I terremoti prima del Mille in Italia e nell’area Mediterranea. SGA-ING, Bologna. Comastri, A., and Traina, G. (1994), Catalogue of Earthquakes in the Mediterranean Region up to the 10th century. Istituto Nazionale di geofisica.

Hancock, P. and Altunel, E. (1997), Faulted archaeological relics at Hierapolis (Pamukkale), Turkey. Journal of Geodynamics 24: 21–36. Hatzfeld, D. (1999), The present-day tectonics of the Aegean as deduced from seismicity, in B. Durand, L. Jolivet, F. Horvath, and M. Seranne (eds.), The Mediterranean Basins: Tertiary Extension within the Alpine Orogen, Special Publication 156. Geological Society, London, 415–26. Besnard, M., Makropoulos, K., Voulgaris, N., Kouskouna, V., Hadjidimitriou, P., Panagiotopoulos, D., Karakaisis, G., Deschamps, A., and Lyon-Caen, H. (1993), Subcrustal microearthquake seismicity and fault-plane solutions beneath the Hellenic Arc. Journal of Geophysical Research 98: 9861–70. Helly, B. (1989), La Grecia antica e I terremoti in E. Guidoboni (ed.), I terremoti prima del Mille in Italia e nell’area mediterranea. SGA, Bologna, 75–90. Hollenstein, C., Kahle, H., Geiger, A., Jenny, S., Goes, S., and Giardini, D. (2003), New GPS constraints on the Africa-Eurasia plate boundary zone in southern Italy. Geophysical Research Letters 30/18: 1935. Hubert, A., King, G. P.C., Armijo, A., Meyer, B., and Papanastasiou, D. (1996), Fault-reactivation, stress interaction and rupture propagation of the 1981 Corinth earthquake sequence. Earth and Planetary Science Letters 142: 573–85. Hunstad, I. and England, P. (1999), An upper bound on the rate of strain in the Central Apennines, Italy, from triangulation measurements between 1869 and 1963. Earth and Planetary Science Letters 169: 261–7. IGME (Institute of Geology and Mineral Exploration) (1989), Seismotectonic Map of Greece, 1:500,000 scale. Jackson, J. A. (1993a), Rates of active deformation in the Eastern Mediterranean, in E. Boschi, A. Morelli, and E. Mantovani (eds.), Recent Evolution and Seismicity of the Mediterranean Region. Kluwer, Dordrecht, 53–64. (1993b), Relations between faulting and continuous deformation on the continents. Annali di Geofisica 36: 3–11. (2001), Living with earthquakes: know your faults. Journal of Earthquake Engineering 5: 5–123. and McKenzie, D. (1988), The relationship between plate motions and seismic moment tensors, and the rates of active deformation in the Mediterranean and Middle East. Geophysical Journal 93: 45–73. Gagnepain, J., Houseman, G., King, G. C. P., Papadimitriou, P., Soufleris, C., Virieux, J. (1982), Seismicity, normal faulting, and the geomorphological development of the Gulf of Corinth (Greece): the Corinth earthquakes of February and March 1981. Earth and Planetary Science Letters 57: 377–97. Jenny, S., Goes, S., Giardini, D., and Kahle, H. (2004), Earthquake recurrence parameters from seismic and geodetic strain rates in the eastern Mediterranean. Geophysical Journal International 157: 1331–47. Jimenez, M.-J., Giardini, D., and Grunthal, G. (2003), The ESC-SESAME Unified hazard model for the EuropeanMediterranean region. EMSC/CSEM Newsletter 19: 2–4; <www.emsc-csem.org/Newsletter/Newsletter_low_19.pdf>, accessed 3 December 2008. Jones, R. E. and Stiros, S. C. (2000), The advent of archaeoseismology in the Mediterranean, in W. J. McGuire, D. R. Griffiths, P. L. Hancock, and I. Stewart (eds.), The Archaeology of Geological Catastrophes, Special Publication 171. Geological Society, London, 25–31.

Earthquakes Karcz, I. and Kafri, U. (1978), Evaluation of supposed archaeoseismic damage in Israel. Journal of Archaeological Science 5: 237–53. and Meshel, Z. (1977), Archaeological evidence for subrecent seismic activity along the Dead Sea-Jordan Rift. Nature 269: 234–5. Kelletat, D. (1991), The 1550 B.P. tectonic event in the Eastern Mediterranean as a basis for assessing the intensity of shore processes. Zeitschrift für Geomorphologie, Suppl. 81: 181–94. Kilian, K. (1996), Earthquakes and archaeological context at Tiryns (Late Helladic Period), in S. Stiros and R. Jones (eds.), Archaeoseismology, Fitch Laboratory Occasional Paper 7. Oxbow, Oxford, 63–8. King, G. and Vita-Finzi, C. (1981), Active folding in the Algerian earthquake of 10 October 1980. Nature 292: 22–6. Kiratzi, A. and Papazachos, C. (1995), Active crustal deformation from the Azores triple junction to the Middle East. Tectonophysics, 243: 1–24. Kirikov, B. (1992), Earthquake Resistance of Structures: From Antiquity to our Times. Mir, Moscow. Kontogianni, V., Tsoulos, N., and Stiros, S. (2002), Coastal uplift, earthquakes and active faulting of Rhodes Island (Aegean Arc): Modeling based on geodetic inversion. Marine Geology 186: 299–317. Koukouvelas, I. and Doutsos, T. (1996), Implications of structural segmentation during earthquakes: the 1995 Egion earthquake, Gulf of Corinth, Greece. Journal of Structural Geology, 18: 1381–8. Lanciani, R. (1918), Segni di terremoti negli edifici di Roma antica. Bolletino della Commissione Archeologica Communale di Roma, 3–30. Le Pichon, X. and Angelier, J. (1979), The Hellenic arc and Trench system: A key to the neotectonic evolution of the Eastern Mediterranean. Tectonophysics 60: 1–42. Chamot-Roke, N., Lallemant, S., Noomen, R., Veis, G. (1994), Cinématique de l’Anatolie-Egée par rapport à l’Europe stable a partir d’une combinaison des mesures de triangulation géodesique sur 80 ans aux mesures de type Satellite Laser Ranging (SLR) récentes. Comptes Rendus de l’Académie des Sciences de Paris 318: 1387–93. McClusky, S. and 27 others (2000), Global Positioning System constraints on plate kinematics and dynamics in the eastern Mediterranean and Caucasus. Journal of Geophysical Research 105: 5695–719. McKenzie, D. (1972), Active tectonics of the Mediterranean Region. Geophysical Journal of the Royal Astronomical Society 30: 109–85. (1978), Active tectonics of the Alpine-Himalayan belt: the Aegean Sea and surrounding regions. Geophysical Journal of the Royal Astronomical Society 55: 217–54. Makropoulos, K. and Burton, P. (1981), A catalogue of seismicity in Greece and adjacent areas. Geophysical Journal of the Royal Astronomical Society 65: 741–62. Mallet, R. (1858), Fourth Report upon the Facts and Theory of Earthquake Phenomena, Report of 28th Meeting of British Association for the Advancement of Science, London. Marco, S., Agnon, A., Ellenbum, R., Eidelman, A., Basson, U., and Boas, A. (1997), 817-year-old walls offset sinistrally 2.1m by the Dead Sea Transform, Israel. Journal of Geodynamics 24: 11–20. Marinatos, S. (1960), Helice: A submerged town of classical Greece. Archaeology 13: 186–93.

489

Meghraoui, M., Jaegy, R., Lammali, K., and Albarede, F. (1988), Late Holocene earthquake sequences on the El Asnam (Algeria) thrust fault. Earth and Planetary Science Letters 90: 187–203. Maouche, S., Chemaa, B., Cakir, Z., Aoudia, A., Harbi, A., Alasset, P., Ayadi, A., Bouhadad, Y., and Benhamouda, F. (2004), Coastal uplift and thrust faulting associated with the Mw 6.8 Zemmouri (Algeria) earthquake of 21 May 2003. Geophysical Research Letters 31/19, art. L19605, 8 October. Meyer, B., Armijo, A., Massonet, D., de Chabalier, J., Delacourt, C., Ruegg, J., Achache, J., Briole, P., and Papanastasiou, D. (1996), The 1995 Grevena (Northern Greece) earthquakes: Fault model constrained with tectonic observations and SAR interferometry. Geophysical Research Letters 23/19: 2677–80. Michetti, A. (1994), Paleosismologa e pericolosita sismica: Stato delle conoscenze ed ipotesi di svilupo. Rendiconto 2, Gruppo nazionale per la difesa dai terremoti. Consiglio Nazionale delle Ricerche, Rome. Brunamonte, F., Serva, L., and Vittori, E. (1996), Trench investigations of the 1915 Fucino earthquake fault scarps (Abruzzo, central Italy): Geological evidence of large historical events, Journal of Geophysical Research 101: 5921–36. Mulargia, F. and Geller, R. (2003), Earthquake Science and Seismic Risk Reduction. NATO Science Series, IV. Earth and Environmental Sciences 32 (plus CR ROM). Nikonov, A. (1996), The disappearance of the ancient towns of Dioscuria and Sebastopolis in Colchis on the Black Sea. A problem in Engineering Geology and Palaeoseismology, in S. Stiros and R. Jones (eds.), Archaeoseismology, Fitch Laboratory Occasional Paper 7. Oxbow, Oxford, 195–204. Nur, A. and Ron, H. (1996), And the walls came tumbling down: Earthquake history in the Holyland, in S. Stiros and R. Jones (eds.), Archaeoseismology, Fitch Laboratory Occasional Paper 7. Oxbow, Oxford, 75–86. and Cline, E. (2000), Poseidon’s horse: Plate tectonics and earthquake storm in the Late Bronze Age Aegean and the Eastern Mediterranean. Journal of Archaeological Science 27: 43–63. Okada, Y. (1985), Surface deformation due to shear and tensile faults in a half space. Bull. Seism. Soc. America 75: 1135–54. Pantosti, D. and Valensise, G. (1989), Riconoscere il ‘terremoto caratteristico’: il caso dell’Appennino centro-meridionale, in E. Guidoboni (ed.), I terrrmoti prima dell’Mille in Italia e nell’area Mediterranea. ING, Bologna, 536–52. Schwartz, D. P., and Valensise, G. (1993), Paleoseismology along the 1980 surface rupture of the Irpinia Fault: implications for earthquake recurrence in the Southern Apennines, Italy, Journal of Geophysical Research 98: 6561–77. Papadimitriou, E., Karakostas, V. G., and Papazachos, B. C. (2001), Rupture zones in the area of the 17.08.99 Izmit (NW Turkey) large earthquake (Mw 7.4) and stress changes caused by its generation. Journal of Seismology 5: 269–76. Papadopoulos, G. and Pavlides, S. (1992), The large 1956 earthquake in the South Aegean: Macroseismic field configuration, faulting and neotectonics of the Amorgos Island. Earth and Planetary Science Letters 113: 383–96. Papaioannou, C. and Papazachos, C. (2000), Time independent and time-dependent seismic hazard in Greece based on seismogenic sources. Bulletin of the Seismological Society of America 90: 22–33.

490

Stathis Stiros

Papastamatiou, D. and Mouyiaris, N. (1986), The Sophades earthquake on 30 April 1954: Field observations by Yannis Papastamatiou [in Greek]. Geology and Geophysics Research Special Issue Athens, 341–62. Papathanassiou, G., Pavlides, S., Christaras, B. and Pitilakis, K. (2005), Liquefaction case histories and empirical relations of earthquake magnitude versus distance from the broader Aegean region. Journal of Geodynamics 40/2–3: 257–78. Papazachos, B., and Papazachou, C. (1997), The Earthquakes of Greece. Zitis, Thessaloniki. Pirazzoli, P. A. (1986), The Early Byzantine tectonic paroxysm. Zeitung fur Geomorphologie, Suppl. NS 62: 31–49. Montaggioni, L. F., Saliege, J., Seconzac, G., Thommeret, Y., and Vergnaud-Grazzini, C. (1989), Crustal block movements from Holocene shorelines: Rhodes island (Greece). Tectonophysics 170: 89–114. Thommeret, J., Thommeret, Y., Laborel, J., and Montagionni, L. (1982), Crustal block movements from Holocene shorelines: Crete and Antikythira (Greece). Tectonophysics 68: 27–43. Laborel, J. and Stiros, S. C. (1996), Earthquake clustering in the Eastern Mediterranean during historical times. Journal of Geophysical Research 101: 6083–97. Stiros, S. C., Fontugne, M., and Arnold, M. (2004), Holocene and Quaternary uplift in the central part of the southern coast of the Corinth Gulf (Greece), Marine Geology 212: 35–44. Robert, L. (1978), Documents d’Asie Mineure. Bulletin de Correspondence Hellenique 102: 395–408. Sachpazi, M., Hirn, A., Nercessian, A., Avedik., F., McBride, J., Loucoyannakis, M., Nicolich., R., and the STREAMERSPROFILES group (1997), A first coincident normal-incidence and wide-angle approach to studying the extending Aegean crust. Tectonophysics 270: 301–12. Saroglu, F., Emre, O., and Kuscu, I. (1992), Active Fault Map of Turkey, 1:1,000,000 Scale, General Directorate of Mineral Research and Exploration (MTA), Ankara. Shaw, B., Ambraseys, N. N., England, P. C., Floyd, M. A., Gorman, G. J., Higham, T. F. G., Jackson, J. A., Nocquet, J. M., Pain, C. C., and Piggott, M. D. (2008), Eastern Mediterranean tectonics and tsunami hazard inferred from the AD 365 earthquake. Nature Geoscience 1: 268–76. Simopoulos, K. (1984), Foreign Travellers in Greece, 1700–1800. 4th edn. Athens, ii [in Greek]. Sinopoli, A. (1991), Dynamic analysis of a stone column excited by a sine wave ground motion. Applied Mechanics Review 44: S246–S255. Soter, S. and Katsonopoulou, D. (1999), Occupation horizons found in the search for the ancient Greek city of Helike. Geoarchaeology: An International Journal 14: 531–63. Stein, R. S. and Yeats, R. S. (1989), Hidden earthquakes. Scientific American 260 (June): 48–57. Barka, A., and Dieterich, J. (1997), Progressive failure on the North Anatolian Fault since 1939 by earthquake stress triggering. Geophysical Journal International 128: 594–604. Stiros, S. C. (1993), Kinematics and deformation of central and southwestern Greece from historical triangulation data and implications for the active tectonics of the Aegean, Tectonophysics 220: 283–300.

(1995), Archaeological evidence of antiseismic constructions in antiquity. Annali di Geofisica 38: 725–36. (1996), Identification of earthquakes from archaeological data: Methodology, criteria and limitations, in Stiros and Jones (1996: 129–52). (1998a), Historical seismicity, palaeoseismicity and seismic risk in Western Macedonia, Northern Greece. Journal of Geodynamics 26: 271–87. (1998b), Archaeological evidence for unusually rapid Holocene uplift rates in an active normal faulting terrain: Roman harbour of Aigeira, Gulf of Corinth, Greece. Geoarchaeology: An International Journal 13: 731–41. (2001), The AD 365 Crete earthquake and possible seismic clustering during the 4–6th centuries AD in the Eastern Mediterranean: a review of historical and archaeological data. Journal of Structural Geology 23: 545–62. (2005), Social and historical impacts of earthquakerelated sea-level changes on ancient (prehistoric to Roman) coastal sites. Zeitschrift für Geomorphologie, Suppl. 137: 79–89. and Drakos, A. (2000), Geodetic constraints on the fault pattern of the 1978, Thessaloniki (Northern Greece), earthquake (Ms = 6.4), Geophysical Journal International 143: 679–88. (2006), A fault-model for the tsunami-associated, magnitude ≥8.5 Eastern Mediterranean, AD 365 earthquake. Zeitschrift für Geomorphologie 146: 125–37. and Jones, R. E. (eds.) (1996), Archaeoseismology, Fitch Laboratory Occasional Paper No. 7. Oxbow, Oxford. and Papageorgiou, S. (2001), Seismicity of Western Crete and the destruction of the town of Kisamos at AD 365: Archaeological evidence. Journal of Seismology 5: 381–97. Pirazzoli, P. A., Laborel, J., and Laborel-Deguen, F. (1994), The 1953 earthquake in Cephalonia (Western Hellenic Arc): coastal uplift and halotectonic faulting. Geophysical Journal International 117: 834–49. Laborel, J., Laborel-Deguen, F., Papageorgiou, S., Evin, J., and Pirazzoli, P. (2000), Seismic coastal uplift in a region of subsidence: Holocene raised shorelines of Samos Island, Aegean Sea, Greece. Marine Geology 170: 41– 58. Stucchi, M. and Camassi, R. (1996), Building up a parametric earthquake catalogue in Europe: the historical background, in D. Giardini (ed.): Proceedings of the NATO Conference on Historic and pre-Historic Earthquakes. Kluwer, Dordrecht. Taymaz, T., Jackson, J., and Westaway, R. (1990), Earthquake mechanisms in the Hellenic Trench near Crete. Geophysical Journal International 102: 695–731. Toksoz, M. N., Shakal, A. F., and Michael, A. J. (1979), Spacetime migration of earthquakes along the North Anatolian Fault Zone and seismic gaps. Pure and Applied Geophysics 117: 1258–70. Turkish General Directorate of Maps (1972), Seismic Zones of Turkey (unpublished). Turkish General Directorate of Maps, Ankara. (1996), Seismic Zones of Turkey, 1:1,800,000 scale. Turkish General Directorate of Maps, Ankara. Valensise, G. and Pantosti, D. (1992), A 125 kyr-long geological record of seismic source repeatability: the Messina Straits (southern Italy), and the 1908 earthquake (Ms 71/2). Terra Nova 4: 472–83.

Earthquakes and Pantosti, D. (2001), The investigation of potential earthquake sources in penisular Italy: A review. Journal of Seismology 5: 287–306. Ward, S. N. and Valensise, G. R. (1989), Fault parameters and slip distribution of the 1915 Avezzano, Italy, earthquake derived form geodetic observations. Bulletin of the Seismological Society of America 79: 690– 710.

491

Westaway, R. and Jackson, J. (1984), Surface faulting in the southern Campania-Basilicata earthquake of 23 November 1980. Nature 312: 436–8. Wright, T. J., Parsons, B. E., Jackson, J. A., Haynes, M., Fielding, E. J., England, P. C., and Clarke, P. J. (1999), Source parameters of the 1 October 1995 Dinar (Turkey), earthquake from interferometry and seismic bodywave modelling. Earth and Planetary Science Letters 172: 23–37.

This chapter should be cited as follows Stiros, S. C. (2009), Earthquakes, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 469–491.

This page intentionally left blank

17

Tsunamis Gerassimos Papadopoulos

Introduction According to Imamura (1937: 123), the term tunami or tsunami is a combination of the Japanese word tu (meaning a port) and nami (a long wave), hence long wave in a harbour. He goes on to say that the meaning might also be defined as a seismic sea-wave since most tsunamis are produced by a sudden dip-slip motion along faults during major earthquakes (Figure 17.1). Other submarine or coastal phenomena, however, such as volcanic eruptions, landslides, and gas escapes, are also known to cause tsunamis. According to Van Dorn (1968), ‘tsunami’ is the Japanese name for the gravity wave system formed in the sea following any large-scale, short-duration disturbance of the free surface. Tsunamis fall under the general classification of long waves. The length of the waves is of the order of several tens or hundreds of kilometres and tsunamis usually consist of a series of waves that approach the coast with periods ranging from 5 to 90 minutes (Murty 1977). Some commonly used terms that describe tsunami wave propagation and inundation are illustrated in Figure 17.2. Because of the active lithospheric plate convergence, the Mediterranean area is geodynamically characterized by significant volcanism and high seismicity as discussed in Chapters 15 and 16 respectively. Furthermore, coastal and submarine landslides are quite frequent and this is partly in response to the steep terrain of much of the basin (Papadopoulos et al. 2007a). Tsunamis are among the most remarkable phenomena associated with earthquakes, volcanic eruptions, and landslides in the Mediterranean basin. Until recently, however, it was widely believed that tsunamis either did not occur in the Mediterranean Sea, or they were so rare that they did not pose a threat to coastal communities. Catastrophic tsunamis are more frequent on

Pacific Ocean coasts where both local and transoceanic tsunamis have been documented (Soloviev 1970). In contrast, large tsunami recurrence in the Mediterranean is of the order of several decades and the memory of tsunamis is short-lived. Most people are only aware of the extreme Late Bronge Age tsunami that has been linked to the powerful eruption of Thera volcano in the south Aegean Sea (Marinatos 1939; Chapter 15). Even that wave is commonly not viewed as a significant geophysical event. Indeed, it is often seen as rather an exotic episode—part myth and part fact—given that it happened in prehistoric times and has been linked with the collapse of the Minoan civilization of Crete. These are some of the reasons why, in comparison to other parts of the world, the scientific study of tsunamis in the Mediterranean Sea has been rather neglected until the last few decades. This chapter reviews the tsunami history of the Mediterranean and evaluates the progress that has been made in assessing the tsunami hazard over its several regions. In addition, the prospects for further development of tsunami science and associated risk mitigation technology in the region are outlined.

Tsunami Quantification A parameter that is of particular importance for understanding tsunami-generating mechanisms and assessing tsunami hazard better is wave size, and this can be expressed as either intensity or magnitude. These parameters, however, are difficult to determine— even for more recent events (e.g. Soloviev 1970; Shuto 1993). In Europe, tsunami intensity (k) is traditionally estimated according to the 6-grade SiebergAmbraseys scale (Ambraseys 1962). In this chapter,

494

Gerassimos Papadopoulos Outer ridge

Trench

Sanriku

1896

(a)

A

Continent

A

1933

Ocean

Intense inelastic deformation

Continental lithosphere

Oceanic lithosphere

(b) Aleutian Continent

1946 1929

Ocean

Stress concentration Sudden interplate faulting

(c)

Tsunamigenic or aseismic complex faulting

Fig. 17.1. Co-seismic dip-slip motion along faults and tsunamigenesis near deep sea trenches. The diagram on the left (modified from Fukao 1979) shows the sequence (a–c) of inelastic deformation and stress concentration culminating in aseismic faulting. In the Sanriku (north-east Japan) and Aleutian examples, thrust faulting is assumed to be the mechanism responsible for the 1896 and 1946 tsunamigenic earthquakes, whilst normal faulting is assumed for the 1933 and 1929 events (modified from Kanamori 1972).

Maximum run-up r

Wave direction Maximum water levell Tsunami height at shore Tsunami

Sea level at time of tsunami Inundation = maximum horizontal intrusion

Fig. 17.2. A schematic explanation of some of the tsunami terms used in this chapter (modified from IOC 1998).

the reported tsunami intensities are based on this scale, while the Murty-Loomis (1980) definition is adopted for tsunami magnitude (ML). An attempt has been made to recalculate the intensities of some important Mediterranean tsunamis according to a recently introduced 12-grade intensity scale (Papadopoulos and Imamura 2001; Papadopoulos 2003a) and these are shown in Table 17.1.

Major Tsunami Events in the Mediterranean In this section the tsunami history of the Mediterranean Sea is reviewed. Emphasis is given to a selection of key tsunami events and these were chosen because of their significant impact on coastal communities and also because they have been extensively studied by a wide range of geophysical, geological, geomorphological, historical, and archaeological research methods as well as by hydrodynamic numerical modelling techniques. This review also aims to evaluate the advantages and disadvantages of the research methods that have been used, to highlight the progress achieved in this area, and to identify the emerging prospects for further research. The tsunami catalogues used in the following review are based mainly on those of Galanopoulos (1960), Ambraseys (1962), Antonopoulos (1979), Papadopoulos and Chalkis (1984), Tinti and Maramai (1996), Soloviev et al. (2000), Papadopoulos (2001 2003b), Tinti et al. (2004) and Fokaefs

Tsunamis TABLE 17.1. Strong tsunamis of intensity k ≥ 4 reported for the Mediterranean Sea between 426 No

Year

Month

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43

426 BC 373 BC 142/4 365 447 551 552 556 749 1169 1202 1303 1343 1365 1389 1402 1481 1609 1612 1627 1650 1693 1741 1748 1759 1766 1773 1783 1817 1823 1856 1866 1866 1866 1867 1908 1944 1948 1956 1963 1979 1999 2002

summer winter

Day

07 01 07 05

21 26 09

01 02 05 08 10 01 03 06 05 04 11 07 10 01 01 05 11 05 05 02 08 03 11 01 02 03 09 12 08 02 07 02 04 08 12

18 04 20 08 18 02 20 03 08 30 11 11 31 25 25 22 06 06 23 05 13 02 02 06 09 28 20 09 09 07 15 17 30

Location

k

Ref

K

Maliakos Gulf W. Gulf of Corinth Rhodes Crete Sea of Marmara Lebanon Maliakos Gulf Kos Levantine coast Messina Straits Syrian coast & Cyprus Crete Sea of Marmara Algiers Chios Gulf of Corinth Rhodes Rhodes Crete Gargano, Italy Thera Eastern Sicily Rhodes W. Gulf of Corinth Akko Sea of Marmara Tangiers Calabria W. Gulf of Corinth North Sicily Chios Albania Kythira Albania Gythion Messina Straits Stromboli Karpathos Cyclades W. Gulf of Corinth Montenegro Sea of Marmara Stromboli

5 5 4 5 4 5 4 4 4 4 4 5 4 4 4 4 4 5 4 4 6 4 5 4 5 4 4 6 4 4 4 4 4 4 4 6 4 4 6 4 4 4 4

P P P P P ** P P ** TM AM P P ** P P ** P P ** P TM P P ** P ** TM P TM P P P P P TM TM P ** P P P *

8 9 7 10 8 8 8 8 7 8 7 10 8 8 6 8 7 8 8 6 10 7 8 9 8 7 7 9 9 8 8 7 6 7 7 10 7 7 9 7 8 6 7

BC

and

AD

495

2002

h (cm)

ML

200

2,000

−1.4 +3.0 +2.3

1,000

900 900 500

−1.8

800

1,300

−0.4

1,500 500

+3.0 −11.0

250 900

Notes: k = tsunami intensity in the 6-grade scale of Sieberg-Ambraseys, K = tsunami intensity in the 12-grade scale of Papadopoulos and Imamura (2001), h = run-up height, ML = Murty-Loomis (1980) tsunami magnitude. * new event. ** revised in this chapter. Sources: AM = Ambraseys (1962), P = Papadopoulos (2001), TM = Tinti and Maramai (1996) and Tinti et al. (2004).

and Papadopoulos (2007). Tsunamigenic zones in the Mediterranean determined from historical data are shown in Figure 17.3a, while Table 17.1 lists the key characteristics of the known strong tsunamis (k ≥ 4). A range of field evidence has been utilized to identify palaeotsunamis and this includes sedimentological, geomorphological, and archaeological data (Figure 17.3b). The following sections review the evidence for tsunami

activity across the Mediterranean region beginning with the Alboran Sea in the west and moving eastwards.

Alboran Sea According to documentary sources, the earthquakes in northern Algeria of 2 January 1365, 6 May 1773, and 21 and 22 August 1856 caused tsunamis of up to

496

Gerassimos Papadopoulos

(a)

Potential Low Intermediate High Very high 2

7 3

15

8 5

16 13 12

4 1

14 11

6

18

9

0

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

10

17

500 km

Alboran Sea Liguria and the Côte d’Azur Tuscany Aeolian Islands Tyrrhenian Calabria Eastern Sicily and Messina Straits Gargano Peninsula East Adriatic Sea Western Hellenic arc Eastern Hellenic arc Cyclades Gulf of Corinth Maliakos Gulf East Aegean Sea North Aegean Sea Sea of Marmara Cyprus Levantine Sea

(b)

Crete Otranto-Leuca

Sarkoy Troy

Itea (3)

Didim Dalaman (3) Thera

Fetiye

Cyprus

Palestine Israel

Archaeological

Geomorphological

Sediment deposits

Pumice

250 km

Fig. 17.3. (a) Tsunamigenic zones in the Mediterranean Sea defined from documentary sources and classified according to their relative tsunami potential as explained in the text. See also Table 17.1. Most of North Africa has not been classified. Here tsunami potential is generally regarded as being low and long-term records are not available. (b) Types of field evidence for the occurrence of past tsunamis.

k 4 (Figure 17.3a). Eyewitness accounts and tide-gauge records verify that similar tsunamis were produced by the strong earthquakes of 9 September 1954, 10 October 1980, and 21 May 2003. In these examples, the location of the earthquakes on land in northern Algeria produced submarine slumping that generated powerful turbidity currents (Soloviev et al. 2000). As a result, atmospheric gravity waves (rissagas) of up to 2-m amplitude and with a frequency of about 10 minutes were observed along the Spanish coast. These waves have also

been observed in southern bays of the Balearic Islands (Monserrat et al. 1991) and also in the Aegean Sea (Papadopoulos 1993a).

Liguria and the Côte d’Azur On 16 October 1979 a submarine slope failure that occurred during the construction of the new Nice airport produced a tsunami wave 3 m high that was observed near Antibes (Figure 17.3a). The near field wave heights

Tsunamis

were successfully simulated by Assier-Rzadkiewicz et al. (2000). The theoretical results were not, however, in complete agreement with the far field observations. This may be explained by the rapid amplitude attenuation of the tsunami due to strong wave dispersion, a common feature of landslide-generated tsunamis (Papadopoulos and Kortekaas 2003; Papadopoulos et al. 2007a). The 20 July 1564 and 23 February 1887 earthquakes (Eva and Rabinovich 1997) triggered tsunamis inundating the coast from Nice to Antibes and from Genoa to Cannes respectively. A modern-day repeat of these intermediate magnitude waves would threaten the densely populated coastal zone of Liguria and the Côte d’Azur.

Tyrrhenian Sea Along the coast of Tuscany in the northern Tyrrhenian Sea (Figure 17.3a), very few events have been reported since the k 4 tsunami of 5 March 1823. On the slopes of Stromboli in the Aeolian Islands (Figures 17.3a and 17.4), volcanic landslides produced tsunamis of k 3 or 4 on 3 July 1916, 22 May 1919, 11 September 1930, 20 August 1944, and 30 December 2002 (Figure 17.5). Further south an extreme event occurred in Tyrrhenian Calabria (Figures 17.3a and 17.4) on 6 February 1783 when a huge earthquake-induced rockfall triggered a k 6 tsunami at the Scilla beach (Tinti and Guidoboni 1988). Inundation heights of 6–9 m were observed and more than 1,500 lives were lost. At Torre del Faro, to the north of Messina, the run-up height was about 6 m and twenty-six people were swept out to sea.

Eastern Sicily and the Messina Straits The 11 January 1693 earthquake in north-east Sicily, that claimed about 70,000 victims, caused a tsunami of k 4. Sea-level oscillations destroyed many boats

Italy Tyrrhenian Sea

South Adriatic Sea

Stromboli 1783 1823

Sicily

1908 Reggio Calabria Messina 1693

1169 100 km

Fig. 17.4. Important tsunamis reported for southern Italy.

497

and ships while flooding was reported from Catania, Augusta, and Messina (Figures 17.3a and 17.4). Tinti et al. (2001) reported hydrodynamical studies of this event and concluded that the faults most likely to be the earthquake source are located in the Scordia-Lentini graben that intercepts the coastline. The 28 December 1908 earthquake is one of the largest ever reported in Italy (Chapter 16). Major towns in southern Italy, including Messina and Reggio Calabria, were completely destroyed with more than 60,000 victims. The earthquake generated a violent tsunami in the Messina Straits (Figures 17.3a and 17.4). This tsunami consisted of at least three large waves that caused many deaths and severe damage to ships, buildings, and property. A tsunami of k 6 was observed along the Calabrian coast at Pellaro and on the Sicilian coast at S. Alessio where wave heights observed were 13 m and 11.7 m respectively. A tsunami with very similar characteristics to the 1908 event is believed to have struck the Messina straits in 1169. A review of the seismological, geological, and geodetic data (Valensise and Pantosti 1992) in association with information on tectonic stress inversion (Neri et al. 2004) indicates that the 1908 earthquake was associated with normal faulting (Chapter 16). However, the seismogenic fault has not yet been identified and this has hindered attempts to develop simulation models of this major Mediterranean tsunami.

Adriatic Sea Along the Adriatic coast of Italy, a tsunami source is related to the seismicity of the Gargano promontory (Figure 17.3a). The destructive earthquake of 30 July 1627, which may have been associated with the Apricena normal fault on land (Patacca and Scandone 2004), caused a k 5 tsunami. Tinti et al. (1995) calculated that large events are expected in this area on average every 228 years. On the eastern side of the Adriatic Sea moderate to strong tsunami events were reported in Albania in 1866 and in Montenegro in 1979 (Figure 17.3a).

The Hellenic Arc The large tsunamigenic earthquake of AD 21 July 365, located off the shore of western Crete (Papadopoulos and Vassilopoulou 2001), is one of the most contentious and debated natural events in Mediterranean Sea history (Chapter 16). The accounts of Marcellinus, Athanasius, and Jerome, which are the closest in time to the event, leave no doubt that a large area was affected since the tsunami propagated to the north-west, west, and south of the Hellenic trench and reached as far as Methoni,

498

Gerassimos Papadopoulos

Fig. 17.5. Some of the damage caused by the Stromboli tsunami of 30 December 2002 (photo: Gerassimos Papadopoulos).

Sicily, Alexandria, and Dalmatia (Guidoboni et al. 1994; Ambraseys et al. 1994; Shaw et al. 2008) (Figures 17.3a and 17.6). Coastal uplift of up to 9 m in western Crete and 3 m in Antikythira (Thommeret et al. 1981; Pirazzoli et al. 1992) have been dated to (calibrated ages) AD 341–439 and 265–491, respectively. Therefore, the 365 event probably corresponds to the dramatic uplift event that raised the harbour at Phalasarna in northwest Crete, by c.6.6 m (Chapter 16). In Phalasarna, tsunami deposits attributed to the AD 365 wave were described by Pirazzoli et al. (1992). According to Dominey-Howes et al. (1998), however, there is no biostratigraphical or lithostratigraphical evidence to infer tsunami sedimentary deposition. This may be due to the 6.6 m of uplift which could have taken place only a few minutes before the wave arrived if the uplift in western Crete and the AD 365 tsunami were caused by the same seismotectonic movement. Shaw et al. (2008) have recently argued that this earthquake took place not on the subduction interface beneath Crete, but on a fault with a dip of c.30◦ within the

overriding plate. Their tsunami propagation calculations produced a damaging tsunami wave throughout the eastern Mediterranean with a repeat time of about 5,000 years for such an event. A range of archaeological, stratigraphical, geomorphological, and radiometric data imply that a tsunami struck Phalasarna after a strong earthquake around AD 66 (Pirazzoli et al. 1992; Dominey-Howes et al. 1998). In more recent times, remarkable earthquake tsunamis were also observed between the Peloponnese and Crete on 9 March 1630, 6 February 1866, and 20 September 1867. Documentary sources indicate that the 8 August 1303 earthquake, which ruptured the eastern Hellenic arc between Crete and Rhodes (Figures 17.3a and 17.6), was one of the largest reported for the historical period in the Mediterranean. The most serious destruction was reported from eastern Crete (Guidoboni and Comastri 1997) where a large tsunami struck Iraklion on the north coast. The sea swept violently into the city with such force that it destroyed buildings and killed inhabitants. In Acre, Israel, people were swept away and

Tsunamis

499

TURKEY

GREECE

Methoni

Dalaman

556

Kos

1609 Fig. 17.8

1866 Antikythira

Rhodes

1481 144 1741

South Aegean Sea Karpathos

1612

Phalasarna Iraklion

365 0

Crete

1948 1303

100 km

Fig. 17.6. Important tsunamis along the Hellenic arc. The Cyclades region (boxed area) is shown in greater detail in Figure 17.8.

drowned by a huge wave and in Alexandria the sea destroyed port facilities. In the easternmost Hellenic arc strong earthquake tsunamis occurred in AD 142 or 144 in Rhodes and Kos, AD 556 in Kos, 3 May 1481, April 1609, 31 January 1741, and 23 May 1851 in Rhodes, and 9 February 1948 in Karpathos (Figures 17.3a and 17.6). Trenches in Holocene sediments at Dalaman, south-west Turkey, have revealed three tsunami sand layers (Figure 17.7) attributed to the 1303, 1481, and 1741 tsunamis respectively (Papadopoulos et al. 2004). The 1609 wave, which violently inundated Rhodes, is missing from the Dalaman stratigraphy. This may be due to the fact that it failed either to penetrate this far inland or to leave a tsunami deposit at the location of the trench site.

The Cyclades The Cyclades has a long record of tsunami events and is a key area for tsunami research (Papadopoulos et al. 2007b). The range of tsunami-generating mechanisms and the rich archaeological and sedimentary record of past tsunami activity mean that it is a key natural laboratory for the investigation of tsunami generation and their impacts. The Minoan eruption is one of the most significant because of its size (Volcanic Explosivity Index = 7),

its possible impact on Late Bronze Age (LBA) civilizations, and the distribution of huge amounts of tephra, thereby creating an important marker horizon around the eastern Mediterranean. The eruption history may have included four main phases (Heiken and McCoy 1984) and concluded with the formation of the caldera that dominates the landscape of Thera (Santorini) (Figures 17.3a and 17.8) (Chapter 15). The most intensive eruption phase lasted for about 3–4 days (Sigurdsson et al. 1990). The Thera event has similarities with Krakatau where the collapse of the volcanogenerated tsunamis that rolled against the shores of Java and Sumatra, with heights up to 35 m, leading to more than 36,400 casualties. From archaeological observations on Amnissos in northern Crete, Marinatos (1939) suggested that the Theran tsunami was linked with the demise of the Minoan civilization. In coastal sites, assemblages of usually rounded pumice, often mixed with seashells, have been attributed to the Minoan tsunami. However, several cases are rather problematic. On Anafi Island, pumice layers at altitudes up to 250 m could not be attributed to Minoan tsunami deposition, as Marinos and Melidonis (1971) suggested, because the pumice there is at least 18,000 years old and of air-borne origin (Keller 1978). Pumice found by Marinatos (1939) in Amnissos was linked to the Theran eruption without any analysis. The

500

Gerassimos Papadopoulos

Naxos

Amorgos Ios

1956 Astypalaea

1650 Thera Kamari

Source of Minoan tsunami (17th Century BC)

Anafi

25 km

Fig. 17.8. Important tsunamis in the Cyclades Islands in the southern Aegean Sea.

Fig. 17.7. An excavated section showing palaeotsunami deposits in Dalaman, south-west Turkey. This section was excavated in 1996 and was approximately 230 m from the present shoreline. Three tsunami sediment layers (dark sediments) are present and these correspond to the 1303, 1481, and 1741 tsunamis in the east Hellenic arc (photo: Gerassimos Papadopoulos).

relationship between the Minoan tsunami and pumice deposits found in coastal sites of Cyprus, Israel, and Palestine is also rather speculative (Francaviglia 1990). Minoura et al. (2000) identified Minoan tsunami deposits in Gouves in northern Crete, and in coastal trenches in Didim and Fethye in south-west Turkey. AMS 14 C dating on fossil shells from Didim and on marine gastropod shells from Fethye, indicated that the eruption may have occurred around the second half of the nineteenth century BC and this places it about 200 years earlier than the previous estimates. In Thera, a 3.5-m thick volcaniclastic deposit intercalated with

third and fourth phase deposits of the Late Bronze Age eruption has been interpreted as tephra reworked by the Minoan tsunami (McCoy and Heiken 2000). Dominey-Howes (2004) found no evidence for any Late Bronze Age tsunami at forty-one coastal sites on Crete and Kos. Futher evidence for the Minoan tsunami comes from the marine sedimentary record. Seismic-reflection surveys in topographic lows of the western Mediterranean and Calabrian Ridges have shown a distinct, acoustically transparent, flat-lying layer, nicknamed ‘homogenites’. These deposits occupy the uppermost part of the sediment column (Kastens and Cita 1981) within a stratigraphic unit characterized by an upward fining grain size that implies deposition in a single event controlled by gravitational settling. Kastens and Cita (1981) calculated that the emplacement occurred between about 4,400 and 3,100 years BP, and that the homogenite was deposited from sediment transport induced by the Minoan tsunami. The thick and structureless homogeneous mud was later recognized in more than fifty gravity cores in a range of contrasting settings (Cita and Aloisi 2000). Mechanisms proposed for the Minoan tsunami include the entry into the sea of both pyroclastic and debris flows propagated in all directions around the island (McCoy and Heiken 2000) along with caldera collapse combined with a large tectonic earthquake (Pararas-Carayannis 1992). To simulate the wave hydrodynamically, Minoura et al. (2000) suggested a sudden volcano collapse, caldera formation, inrush of water into the caldera, and collision of water masses

Tsunamis

501

Fig. 17.9. Palaeotsunami investigation within an archaeological excavation in St George, Thalassitis, Kamari, in eastern Thera. The box indicates a distinctive sediment layer produced by a tsunami and highlights the area shown in more detail in Figure 17.10 (photo: Gerassimos Papadopoulos.)

with the caldera wall. Their simulation resulted in wave heights of >15 m in the near-field zone and of 6–11 m in northern Crete. However, the extent of the wave inundation was only several hundred metres and, although the fishing and trading economy could have been affected by the destruction of boats and harbour installations, it can be argued that a tsunami of this size would have had little long-term influence on the Minoan civilization. Another large tsunami was generated during the eruption of Columbo, a submarine volcanic edifice lying 7.3 km to the north-east of Thera (Figure 17.8). The main volcanic activity began on 26th September 1650, while volcanism on 30 September was followed by a pause in activity. During this pause a sea swell encircled the whole of Thera island and the tsunami inundated the eastern coast and swept away churches, enclosures, boats, trees, and agricultural land. On the east and west coast of Patmos island and on Ios island,

tsunami run-up heights of 30, 50, and 16 m respectively were reported. Ships and fishing boats moored at Iraklion were swept violently offshore, while vessels were crushed when the wave overtopped the city walls. The volcanic and seismic quiescence that prevailed before the tsunami struck implies that it was generated by submarine landsliding or collapse of the volcanic cone rather than by a strong earthquake or volcanic explosion (Dominey-Howes et al. 2000). A geological record of the 1650 tsunami has been recognized from the coastal site of St George’s near Kamari village in eastern Thera (Figures 17.9 and 17.10). However, Dominey-Howes et al. (2000) were unable to trace any tsunami deposit signature across three trenches, one of them being only 500 m from the St George’site. The non-volcanic clasts they found were angular to very angular in form and it is therefore thought that the non-volcanic sediments may reflect local colluvial processes. Thus, alternative hypotheses involving discontinuous sediment deposition

502

Gerassimos Papadopoulos

sides of the Amorgos basin as well as sea floor sediment instability and a geologically very recent slump (24 × 6 km in area and 3.6 × 106 m3 in volume) occupying part of the basin (Perissoratis and Papadopoulos 1999). The proximity of the slumped area to the earthquake epicentre implies seismic ground accelerations much higher than the minimum values required to initiate slumping. The slump episode may have occurred in association with the 1956 earthquake. Numerical simulations showed a discrepancy by a factor of 3–10 between the maximum reported (30 m) and simulated (3–10 m) wave amplitudes at the source region (e.g. Yalçiner et al. 1993). Field observations and interviews with eyewitnesses (Dominey-Howes 1996; Papadopoulos et al. 2005) showed that the wave amplitude may not have exceeded about 15 m in Astypalaea. Therefore, part of the discrepancy could be explained by an overestimation of the initially reported wave height. Nonetheless, a significant discrepancy remains unexplained. Thus it may be concluded that an adequate reproduction of the near-field wave amplitudes requires not only co-seismic sea-floor fault displacement but also an additional tsunamigenic component such as that from a co-seismic, massive submarine sediment slump.

Fig. 17.10. Detail of the tsunami deposits exposed in the section shown in Figure 17.9 and attributed to the 30 September 1650 event. The tsunami deposits are the dark layer behind the pen in the middle of the photograph (photo: Gerassimos Papadopoulos).

The Gulf of Corinth

and overestimation of the event’s magnitude have been considered (ibid.). The most recent large Mediterranean tsunami occurred in the Cyclades (Figure 17.3a) on 9 July 1956 after a Ms 7.4 crustal earthquake associated with normal faulting (Papadopoulos and Pavlides 1992; Chapter 16). The tsunami-generating rupture was about 100 km in length within the NE–SW trending basin formed by the Thera, Amorgos, and Astypalaea islands (Papazachos et al. 1985). Initial estimates of the near-source wave height varied between 15 and 30 m in Amorgos and Astypalaea (Galanopoulos 1957; Ambraseys 1960) (Figure 17.8). Four people drowned and extensive destruction was noted in port facilities and in both small and large vessels as well as on cultivated land and other property. Two interrelated problems arise out of the study of the 1956 tsunami. The first is the generation mechanism and the second is that numerical simulations have failed to reproduce the observed wave heights accurately. From tide gauge records Galanopoulos (1957) and Ambraseys (1960) concluded that the wave was probably produced by co-seismic landslides. Submarine geophysical survey showed normal faulting in the

The Gulf of Corinth is especially prone to tsunamis due to high seismicity, steep bathymetry and a susceptibility to coastal landsliding (Papadopoulos 2003b; Papadopoulos 2007a). It is shown on Figure 17.3a as having the highest tsunami potential in the Mediterranean region. In 373 BC it has been argued that the town of Helike, located about 7 km east of modern Aeghion (Figure 17.11a), was destroyed by an earthquake and tsunami (Guidoboni et al. 1994; Papadopoulos 1998; and see Chapter 16 for further discussion). Ten Spartan ships at anchor close by were destroyed. The lethal earthquakes of 25 May 1748 and 23 August 1817 generated similar tsunamis in Aeghion causing human losses and extensive damage to vessels, port facilities, and cultivated land. The June 1402 tsunami was also of high intensity and followed a large, possibly near-shore earthquake (Papadopoulos et al. 2000). Interestingly, the near-shore earthquake of 26 December 1861, having a magnitude comparable to the previous events, produced a tsunami of much lower intensity. Seismically triggered earth slides caused local tsunamis along the north coast on 11 June 1794, 6 July 1965, 11 February 1984, and 15 June 1995. An aseismic tsunami generated by sediment slumping at

Tsunamis

503

50 km

(a) 426 BC

M

ali

ak os

GREECE

Gu

lf

552 1817 1748,1963 Aeghion

373 BC 1402

Athens Corinth

Peloponnese

(b)

Fig. 17.11. (a) Important tsunamis in the Gulf of Corinth and the Maliakos Gulf. (b) An air photograph showing the coast to the east of the village of Psathopyrgos prior to the large landslide that produced a tsunami in the west of the Gulf of Corinth on 7 February 1963 (after Galanopoulos et al. 1964). The black line marks the extent of the sediment mass that slipped seawards. This body of sediment was approximately 1500 m in length with a maximum width of 350 m.

a river mouth hit both coasts at the western end of the Gulf of Corinth on 7 February 1963 (Figure 17.11b). The wave killed two people, injured twelve, and was responsible for serious damage to houses, cultivated land, and fishing boats (Galanopoulos et al. 1964). Numerical modelling results are consistent with run-up and inundation observations (Koutitas and Papadopoulos 1998). A similar wave of lesser intensity was observed near Aeghion on 1 January 1996.

The Maliakos Gulf The Maliakos Gulf is located on the western side of the Aegean and strong earthquake-generated tsunamis have been reported here for 426 BC and AD 551 or

552 from classical sources and archaeological evidence (Figures 17.3a and 17.11a). Records of past tsunami activity are not always consistent and care must be taken in their interpretation. For example, Papaioannou et al. (2004) suggested that the 426 BC event was, in fact, rather moderate and they argue that the large tsunami from this period may have occurred during the third century BC and that previous researchers amalgamated the two events into the earlier one at 426 BC. The Byzantine writer Procopius reported on an earthquake that struck the Gulf of Corinth in AD 551. He also described a strong tsunami in the Maliakos Gulf. However, it seems that Procopius did not actually describe a tsunami caused by the AD 552 earthquake, instead he just reproduced the classical sources detailing the 426 BC tsunami.

504

Gerassimos Papadopoulos

East and North Aegean Sea A distinctive tsunami-prone region is associated with an earthquake zone around Chios Island in the eastern Aegean Sea (Figure 17.12a). Tsunamis of k 3 or 4 were observed on 20 March 1389, 12 May 1852, 8

(a) 1932

1893 Ierissos

Samothraki ~ 1300 BC

Troy

TURKEY

1856

Chios

Bosphoru s

50 km

(b)

Bosphorus

TURKEY Tekirdag

Istanbul

1343 447

1912

1999 Bay Izmit

1766

September 1852, 13 November 1856, 2 February 1866, 3 April 1881, and 23 July 1949. Based on the writings of the Greek historian Democles (fourth century BC), some catalogues list a tsunami that supposedly struck Troy in about 1300 BC. A relevant passage in the Wall War of Iliad has not attracted much attention to date. In fact, rhapsody M18–19 describes how the wall of Greeks was damaged by river flooding, yet in M26–33 it becomes clear that the sea contributed to this destruction. In fact, the wall was inundated by the sea and trees and wall materials were carried away by the water; the ground was levelled by the strong wave and the coastal zone was covered by sand; while the sea caused the flow of rivers to reverse. The vivid descriptions in the Iliad indicate that a palaeotsunami survey along the coast of Troy could potentially be of great geomorphological and archaeological interest. In the northern Aegean Sea (Figures 17.3a and 17.12a), a destructive wave that reportedly struck Potidaea on the Chalkidiki Peninsula, in April 479 BC, may imply tsunami action. This is the first tsunami described in documentary sources anywhere in the world. In AD 544 an earthquake-induced destructive inundation hit the coast of Thrace. More recently, seismic tsunamis of k 3 were observed in this area on Samothraki Island on 9 February 1893 and at Ierissos, Chalkidiki, on 26 September 1932.

Sea of Marmara

The Sea of Marmara 50 km

(c) CYPRUS

551

LEBANON

1759 1202 Akko

100 km

749

ISRAEL

JORDAN

Fig. 17.12. Important tsunamis in (a) the east and north Aegean Sea, (b) the Sea of Marmara, and (c) the Cyprus–Levantine Sea area.

This area has a long record of tsunami activity and k 3 or 4 tsunamis were caused by earthquakes on AD 26 January 447, 26 October 740, 2 September 1754, and 19 April 1878 in Izmit Bay. Tsunamis of k 3 or 4 were also caused by earthquakes on AD 25 September 478, 10 August 1265, 18 October 1343, 25 May 1419, 10 September 1509, and 10 July 1894 in Constantinople (Istanbul) as well as on 22 May 1766 in the Bosporus Straits, and on 9 August 1912 in Terkirda˘g (Figures 17.3a and 17.12b). Due to higher seismicity, the east is the most tsunami-prone side of the Marmara Sea. The large (Mw 7.4) Izmit earthquake of 17 August 1999, caused by right-lateral strike-slip faulting with a significant normal component, generated damaging waves up to 2.5 m high in De˘girmendere on the south coast, and elsewhere (Yalciner et al. 2002) with intensities up to 4 (Papadopoulos 2001). Rothaus et al. (2004) argued that the uniformity of the tsunami impact, indicating a wave coming to the south coast from 310◦ , suggests that submarine faulting was the major source of these tsunamis.

Tsunamis

Cyprus-Levantine Sea The final part of this review looks at the coasts bordering the Cyprus and Levantine Sea in the eastern Mediterranean. Archaeological excavations in Kourion, south-west Cyprus (Figures 17.3a and 17.12c), have revealed a destruction horizon attributed to the AD 21 July 365 earthquake tsunami that supposedly devastated Kourion (Soren 1988). However, an earthquake location off the shore of south-west Cyprus is not consistent with both the AD 365 seismic damage distribution and tsunami wave propagation to the south, west, and north-west of Crete (Figure 17.6). An alternative explanation has been put forward that Kourion was hit by a non-tsunami-generating earthquake around AD 370 (Ambraseys 1965; Guidoboni et al. 1994). In Cyprus, earthquake tsunamis were reported on 22 May 1201 or 1202, 11 May 1222, and 10 September 1953, the first two being of high intensity. Geomorphic evidence along coastal sections of southern Cyprus and radiocarbon dating results indicate tsunami activity between 1530 and 1821 (Whelan and Kelletat 2002). In the left-lateral strike-slip Levantine rift, tsunamigenerating earthquakes have been identified from historical records (Figures 17.3a and 17.12c). On 9 July 551 the sea retreated for a mile and many ships were destroyed along the coasts of Lebanon, Syria, and Palestine, while after an earthquake on AD 18 January 749, the waves ‘rose up to the sky’ and destroyed most of the cities and villages along the coasts of Israel, Palestine, and Syria (Guidoboni et al. 1994, Darawcheh et al. 2000). Similar tsunamis were reported after strong earthquakes on 5 December 1033 (that shook the region around the Jordan Valley), on 29 May 1068 (an event possibly centred to the south of the Levantine rift), and on 14 January 1546 near the Dead Sea (Ambraseys et al. 1994).

Tsunami Generation Mechanisms The large propagation areas of waves such as those of AD 365 and 1303 events leave little doubt that they were produced by co-seismic, sea-bottom dip-slip faulting. However, with only limited data about other tsunamis associated with large earthquakes of dip-slip faulting—such as the 1908 event in the Messina straits and the 1956 event in the Cyclades—it is often difficult to distinguish between co-seismic, landslide, or combined source mechanisms. In addition, the generation of tsunamis with strong local effects by earthquakes occurring on land—such as along the strike-slip Levantine rift—remains unexplained. Submarine landsliding

505

triggered by powerful seismic shaking is a possibility, but an alternative mechanism is the dynamic excitation of tsunamis due to seismic energy transmission to the continental shelf by the vertical component of the longperiod Rayleigh waves. However, in the highly seismogenic Cephalonia–Lefkada system of strike-slip faulting in the Ionian Sea, the tsunami activity is low—and in the Sea of Marmara, the tsunami potential is only intermediate (Figure 17.3a). This may be due to the fact that the earthquake focal mechanism involves mainly a strike-slip rather than a dip-slip component, while earth slumping may contribute to more local tsunami generation. Volcanic eruptions are much less frequent than earthquakes in the Mediterranean and only a few tsunamis are therefore attributed to volcanic activity (Chapter 15). However, two of the largest known tsunamis were generated by strong eruptions in the Thera volcanic complex—namely the seventeenth-century BC Minoan tsunami and the AD 1650 Columbo tsunami. The generation of the Minoan tsunami may have included several mechanisms including caldera collapse, strong earthquake activity, pyroclastic surges and flows, and lahars and debris flows into the sea. Additional mechanisms that are worthy of more research include tsunami generation from atmospheric pressure changes due to the volcanic eruption, as well as abrupt sea-bottom impulse following the caldera collapse. The Columbo tsunami was probably caused by the partial collapse of the submarine caldera. Local but strong tsunamis caused by landsliding during eruptive activity of Stromboli have been reported repeatedly (Chapters 1 and 15). Aseismic submarine or coastal landsliding is an important agent of locally powerful tsunamis (Papadopoulos et al. 2007a). The physiographic features of the Gulf of Corinth favour these processes and the most recent events were observed in 1963 and 1996. In the western Mediterranean in 1979, the collapse of a mass of unconsolidated artificial embankment into the sea in Nice on the Côte d’Azur, caused a local tsunami of similar character to that of 1963 in the Gulf of Corinth. Both events produced a large amplitude wave leading to loss of life and significant destruction in the near-source coastal zone.

Tsunami Hazard This section evaluates tsunami hazard in Mediterranean coastal regions using four independent approaches. The first is a comparative description of the spatial distribution of tsunami events based on the geological

Tsunami frequency (n)

506

Gerassimos Papadopoulos

30

70

25

60 50

20 40 15 30 10

20

5

10 0

0 3

W

0

5

10

15

20

Longitude (°)

25

30

32

35

35

40

45

47

Latitude (°N)

E

Fig. 17.13. The frequency (n) of tsunamis in the Mediterranean Sea as a function of longitude (left) and latitude (right) for the period 1628 BC to AD 2003.

and historical tsunami record over the last 3.5 millennia. Second, tsunami recurrence is calculated from empirical intensity–frequency relationships. The third is a statistical approach to estimate tsunami potential in a uniform way across the several tsunamigenic zones of the Mediterranean Sea. Finally, the occurrence of tsunamis is examined as a function of earthquake size. The data sets used here come from the various tsunami catalogues listed in the earlier sections. The known tsunami sources in the Mediterranean are concentrated in eighteen zones of high seismicity and/or volcanism as shown in Figure 17.3a (Chapters 15 and 16). In the Mediterranean as a whole, the tsunami hazard (in terms of the event count) increases gradually from west to east—again tracking the basin-wide pattern of seismicity—although a decrease is apparent in the easternmost part of the basin. In fact, maximum activity occurs between 20◦ E and 30◦ E in the complex tectonic terrain and highly seismically-active structures of Greece (Figure 17.13) including the western and eastern segments of the Hellenic arc and the Gulf of Corinth. Significant tsunami activity is also concentrated in the seismic and volcanic region of Italy between 15◦ E and 17◦ E. The main belt of tsunami activity lies within a latitudinal zone between approximately 37◦ N and 41◦ N, which includes Greece, the Marmara Sea, the Tyrrhenian Sea, and southern Italy (Figure 17.13). To determine the frequency of tsunami occurrence, some authors have used the intensity–frequency relationship: log Nc = ac − bk

(1)

where Nc is the cumulative number of events of intensity equal to or larger than k observed in a time interval of c years, and a c and b are parameters determined by the data. Equation (1) is analogous to the classic earthquake magnitude–frequency relationship (Gutenberg and Richter 1944) and applies to tsunamiprone regions where relatively good datasets are available. For c = 1 formula (1) becomes: log N = a − bk

(2)

a = a c − log c

(3)

where:

Then, the mean tsunami recurrence for intensity ≥k is equal to: T = 10b k−a

(4)

The most likely maximum tsunami intensity (kt ) expected in a time interval of t years is given by: kt = (a + log t)/b

(5)

However, the small number of data points does not allow the application of this relationship in most of the tsunami-prone areas of the Mediterranean. Thus, this approach may only be applied in the larger tsunamigenic regions such as Greece, Italy, and the entire Mediterranean Sea. From the temporal distribution of the reported tsunami events a complete record for events of kc ≥ 3 from AD 1600 onwards is assumed (Figure 17.14). The results are shown in Figure 17.15 and Table 17.2 and suggest that, in the entire Mediterranean, the mean recurrence interval for tsunamis of

Tsunamis 300

2.5

(a)

(a) Greece

250

2.0

n =13.869e

200

log frequency (n)

Tsunami frequency (n)

507

0.001t

2

R= 0.938

150 100

1.5 log n = 3.67–0.61k

1.0

r2 = 0.996 0.5

50

0 0

0 2000

1000

1000

0

1

2

3

4

5

6

2.5

3000

2000

(b) Italy

Years (BC/AD)

log frequency (n)

2.0 6

Tsunami intensity (k)

(b) 5 4

1.5 1.0 log n = 2.16–0.31k

0.5

3

r2 = 0.972 0

2

0

1

2

3

4

5

6

2.5 1000

500

500

0

1000

(c) Mediterranean

2000

1500

Years (BC/AD)

Fig. 17.14. (a) The cumulative frequency (n) of Mediterranean Sea tsunamis as a function of time for the period 1628 BC to AD 2003. (b) The intensity (k) of tsunamis between 1500 BC and AD 2000.

2.0

log frequency (n)

1 1500

1.5 log n = 3.56–0.51k

1.0

r2 = 0.999 0.5

kc ≥ 3, 4, 5, and 6 is 4, 12, 40, and 130 years, respectively. This analysis also shows that Greece is characterized by a higher tsunami frequency than Italy, with the exception of k 6 events, which appear to recur more frequently in the Italian region. This last outcome, however, should be treated with some caution since it is based only on the post-1600 statistics and some very significant high-intensity events, like the AD 365 and 1303 tsunamis in the Hellenic arc and the 1402 event in the Gulf of Corinth, were not considered in the calculation. TABLE 17.2. Mean return period (T ) of tsunami intensity (k) and the most likely maximum tsunami intensity (kt ) to be observed in time interval t for various parts of the Mediterranean Sea Region

Mediterranean Sea Greece Italy Gulf of Corinth*

T (years)

kt

k≥3

4

5

6

t (yrs) = 1

10

100

4 6 26 40

13 24 55 103

41 98 115 261

132 399 242 662

2 2

4 4 2 2

6 5 5 4

* Modified from Papadopoulos (2003b).

0 0

1

2

3

4

5

6

Tsunami intensity (k)

Fig. 17.15. Diagrams showing the relationship between intensity and frequency of tsunamis from the database for (a) Greece, (b) Italy, and (c) the Mediterranean Sea.

Since detailed intensity-frequency relationships are impossible to obtain in most tsunami-prone regions due to small datasets, an alternative procedure was introduced for the uniform assessment of tsunami potential. This procedure combines information on the frequency and intensity of events, but does not consider the potential hazard to far-field locations which is due to wave propagation in remote places. Thus, the tsunami potential or hazard (H ) of a particular zone or area is defined as the normalized quantity: H = Ha /Hmi n

(6)

where the absolute potential, Ha , is an increasing function of the weighted event intensity and the event

508

Gerassimos Papadopoulos

frequency: Ha =

n 

(kci × jc )/tc

i =1

where Hmi n = min {Ha }, kci is the intensity of event i, and jc is a weighted factor of kc . Factor jc has been defined to follow a power law of base 2:21 for kc = 3, 22 for kc = 4, 23 for kc = 5, and 24 for kc = 6. Only events of kc ≥ 3 were considered. The lower date of the time interval (tc ) over which the data set is complete is variable, while the upper date was fixed at the end of 2003. The earliest event in the record was assumed to be AD 300 for kc = 6, AD 1000 for kc = 5, AD 1600 for kc = 4, and AD 1750 for kc = 3 (Figure 17.14). Equation (6) produces a relative tsunami potential scale with a minimum value equal to 1. The eighteen tsunamigenic zones in the Mediterranean were classified into four groups of potential and these are shown in Figure 17.3a and Table 17.3. The lowest potential was found in Tuscany (H = 1), while low potential characterizes the coastal region from Liguria to the Côte d’Azur and the north Aegean Sea. The potential may be described as intermediate in the Alboran Sea, the Aeolian Islands, the Tyrrhenian Calabria, eastern Sicily, and the Messina Straits, the Gargano promontory, the coasts of Albania and Montenegro, the Cyclades, the eastern Aegean Sea, the Sea of Marmara, and the Levantine Sea. High potential

TABLE 17.3. Tsunami potential in each of the tsunamigenic zones of the Mediterranean shown in Figure 17.3a No in Fig. 17.3a 3 7 2 15 6 11 8 16 5 18 1 4 14 9 10 12

Region Tuscany Gargano, Italy Liguria and the Côte d’Azur North Aegean Sea East Sicily and Messina Straits Cyclades Islands Eastern Adriatic Sea Sea of Marmara Tyrrhenian Calabria Levantine Sea Alboran Sea Aeolian Islands Eastern Aegean Sea Western Hellenic arc Eastern Hellenic arc Gulf of Corinth

Ha

H

0.024 1 0.040 1.67 0.047 1.96 0.047 1.96 0.096 4.0 0.113 4.71 0.119 4.96 0.126 5.25 0.127 5.29 0.127 5.29 0.134 5.58 0.134 5.58 0.142 5.92 0.199 8.29 0.223 9.29 0.308 12.83

Potential low low low low intermediate intermediate intermediate intermediate intermediate intermediate intermediate intermediate intermediate high high very high

Note: Ha = absolute potential, H = normalized potential. Very low potential has been tentatively assigned to Campania, Ionian Apulia and Calabria, the central and northern Ionian Sea, the Maliakos Gulf, and Cyprus. See text for further discussion.

values were determined for the western and eastern segments of the Hellenic arc, while the Gulf of Corinth is classified as having the highest tsunami potential (H = 12.83) in the Mediterranean Sea (Figure 17.3a; Chapter 13). In such regions as Cyprus, the Maliakos Gulf, the central and northern Ionian Sea, as well as Ionian Apulia and Calabria, only a few strong or moderate tsunamis have been documented—but these took place before the time period of complete data sets used in the above analysis. A characteristic example is the tsunami wave of intensity k 4 of AD 1222 in Cyprus (Fokaefs and Papadopoulos 2007). In these regions the tsunami potential is not negligible, but it is impossible to quantify at present using the methods outlined above. Thus, these regions are provisionally classified as being of very low tsunami potential. The tsunami database for much of the North African coast of Tunisia, Libya, and Egypt is also very patchy and this probably reflects a combination of low tsunami potential (due to lower relief and low seismicity) and limited documentary and archaeological records. It is important to appreciate that the intensity of an individual tsunami depends on many factors including earthquake size and epicentral location as well as the nature of the crustal displacement (Lorito et al. 2008) The local and regional sea floor bathymetry is also important—as is the geomorphology of the coastal zone receiving the tsunami wave. Given the range of variables, it is not surprising that there is not a strong correlation between tsunami intensity and earthquake magnitude and intensity (Figure 17.16). However, it is possible to identify an upper bounding envelope in both diagrams which shows that tsunami intensity does not exceed a certain value unless it is generated by an earthquake that exceeds a minimum size. For example, to produce a tsunami with an intensity ≥ 4 requires at least an earthquake magnitude of 6.3 or an earthquake intensity of 7.

Risk Mitigation Technology Tsunami risk mitigation can be achieved through a series of actions including: 1. the operation of instrumental early warning systems; 2. the construction of breakwaters; 3. the development of risk mapping using GIS tools; 4. the implementation of dedicated civil protection plans and public education.

Tsunamis

(a)

(b) 6

6

0.41M

0.36i

k = 0.337e r2 = 0.999

k = 0.313e

5

2

Tsunami intensity (k)

5

Tsunami intensity (k)

509

4 3 2 1

r = 0.996 4 3 2 1 0

0

IV

V

VI

VII

VIII

IX

X

XI

XII

Earthquake intensity (i )

5

6

7

8

Earthquake magnitude (M)

Fig. 17.16. Tsunami intensity (k) as a function of (a) earthquake intensity and (b) earthquake magnitude for the entire Mediterranean Sea database. Upper bounding envelopes are shown.

Although the countries bordering the Mediterranean Sea do not have an established tradition in this field (in contrast to Japan, for example, and other circumPacific countries), significant progress has been made over the last fifteen years or so. For example, experimental tsunami warning systems, consisting of both seismic and tide-gauge instruments, have recently been tested in the Messina Straits and the southern Aegean Sea (Piscini et al. 1998, Papadopoulos 2003c). After the strong tsunami of 30 December 2002 in Stromboli, the civil protection authorities established a local tsunami alarm system. A major problem, however, for tsunami risk mitigation in the Mediterranean Sea is that travel times from tsunami sources to threatened coastal communities are short, ranging from a few minutes to around one hour depending on the source location. However, operationally viable systems can be achieved by minimizing the alarm time to no more than about five minutes from the earthquake origin time. In 2005 the Intergovernmental Oceanographic Commission of UNESCO initiated a project of establish an instrumental early warning system for tsunamis in the Mediterranean Sea and the northeast Atlantic Ocean. Pilot studies of risk mapping on a microzonation basis appear to be useful for long-term planning in the mitigation of tsunami risk (Papadopoulos and Dermentzopoulos 1998; Papathoma et al. 2003). In Japan, the construction of breakwaters has been used for the protection of some coastal zones from tsunami attack, but such measures have never been recommended in the Mediterranean Sea. In comparison to other natural hazards in the Mediterranean such as earthquakes,

volcanoes, and large fluvial floods, there has only been limited progress to date in the development of civil protection plans and raising public awareness of the potential threat from tsunamis in the region. While public awareness of the tsunami hazard more generally has undoubtedly been raised following the catastrophic tsunami in the Indian Ocean in December 2004, awareness of the extent of the local tsunami threat in the Mediterranean region is still rather limited. Progress in the types of action outlined above is very important for the development of an effective tsunami risk mitigation policy for the Mediterranean Sea coasts. Using tsunami-wave simulations generated from mathematical models, Lorito et al. (2008) have calculated a range of parameters including wave height profiles and wave travel times to assess the potential threats to southern Italy from tsunamis generated by earthquakes in three different source zones, namely the southern Tyrrhenian thrust belt, the Tell-Atlas thrust belt, and the western Hellenic arc. The outcomes show a highly variable impact for tsunamis produced in the different source zones and this analysis highlights the need for scenario testing to be undertaken at the scale of the entire Mediterranean basin (Lorito et al. 2008).

Concluding Remarks Segments of the Mediterranean Sea coast have been struck in the past by large, destructive tsunamis generated by submarine earthquakes and volcanic eruptions (e.g. Soloviev et al. 2000), while powerful tsunamis

510

Gerassimos Papadopoulos

have also been locally generated by landslides (e.g. Papadopoulos 1993b). The recurrence interval for tsunami intensities ≥ 4, 5, and 6 is of the order of 12, 40, and 130 years respectively. In this chapter, the highest tsunami potential has been calculated in the Hellenic arc and the Gulf of Corinth. However, infrequent but large events have been recognized for the Cyclades in the southern Aegean Sea and the Straits of Messina in southern Italy, as well as in the Alboran, Levantine, and Marmara Seas. It is clear, therefore, that tsunami waves should not be neglected as a potential source of risk that threaten coastal communities of the Mediterranean. Several approaches offer considerable potential to improve our understanding of the tsunami phenomena and associated hazards including the identification of earthquake, volcanic eruption, and landslide sources and mechanisms as well as palaeotsunami investigations. To improve tsunami risk assessment, more work is needed to extend the event database, to simulate wave generation, propagation, and coastal inundation using numerical models and to map the components of physical and anthropogenic risk with the use of GIS tools. Recent events in south-east Asia and the Indian Ocean margins have highlighted the importance of developing risk mitigation strategies that include public awareness activities, the development of instrumental tsunami warning systems and the elaboration of specific civil protection plans across the Mediterranean region.

References Ambraseys, N. N. (1960), The seismic sea wave of July 9, 1956, in the Greek Archipelago. Journal of Geophysical Research 65: 1257–65. (1962), Data for the investigation of the seismic sea-waves in the eastern Mediterranean. Bulletin of the Seismological Society of America 52: 895–913. (1965), The seismic history of Cyprus. Revue pour l’étude des Calamites 40: 25–48. Melville, C. P., and Adams, R. D. (1994), The Seismicity of Egypt, Arabia and the Red Sea: A Historical Review. Cambridge University Press, Cambridge. Antonopoulos, J. (1979), Catalogue of tsunamis in the eastern Mediterranean from antiquity to present times. Annali di Geofisica 32: 113–30. Assier-Rzadkiewicz, S., Heinrich, P., Sabatier, P.C., Sovoye, B., and Bourillet, J. F. (2000), Numerical modelling of a landslidegenerated tsunami—the 1979 Nice event. Pure and Applied Geophysics 157: 1707—27. Cita, M. B. and Aloisi, G. (2000), Deep-sea tsunami deposits triggered by the explosion of Santorini (3500 years BP), eastern Mediterranean. Sedimentary Geology 135: 181–203. Darawcheh, R., Sbeinati, M. R., Margottini, C., and Paolini, S. (2000), The 9 July 551 AD Beirut earthquake, Eastern Mediterranean region. Journal of Earthquake Engineering 4: 403–14.

Dominey-Howes, D. (1996), Sedimentary deposits associated with the July 9th 1956 Aegean Sea tsunami. Physics and Chemistry of the Earth 21: 51–5. (2004), A re-analysis of the Late Bronge Age eruption and tsunami of Santorini, Greece, and the implications for the volcano-tsunami hazard. Journal of Volcanology and Geothermal Research 130: 107–32. Dawson, A., and Smith, D. (1998), Late Holocene coastal tectonics at Falasarna, W. Crete: a sedimentary study, in C. Vita-Finzi (ed.), Coastal Tectonics, Special Publication 146. Geological Society, London, 343–52. Papadopoulos, G. A., and Dawson, A.G. (2000), Geological and historical investigation of the 1650 Mt. Columbo (Thera Island) eruption and tsunami, Aegean Sea, Greece. Natural Hazards 21: 83–96. Eva, C. and Rabinovich, A. B. (1997), The February 23, 1887 tsunami recorded on the Ligurian coast, western Mediterranean. Geophysical Research Letters 24: 2211–14. Fokaefs, A. and Papadopoulos, G. A. (2007), Tsunami hazard in the Eastern Mediterranean: strong earthquakes and tsunamis in Cyprus and the Levantine Sea. Natural Hazards 40: 503–26. Francaviglia, V. (1990), Sea-borne pumice deposits of archaeological interest on Aegean and eastern Mediterranean beaches, in D. A. Hardy et al. (eds.), Thera and the Aegean World III. Thera Foundation, London, iii. 127–34. Fukao, Y. (1979), Tsunami earthquakes and subduction processes near deep-sea trenches. Journal of Geophysical Research 84: 2303–14. Galanopoulos, A. G. (1957), The seismic sea wave of July 9, 1956. Praktika Akadimias Athinon 32: 90–101. (1960), Tsunamis observed on the coasts of Greece from Antiquity to present time. Ann. di Geof. 13: 369–86. Delibasis, N., and P. Comninakis, P. (1964), A tsunami generated by an earth slump set in motion without shock. Annales Geologic Pays Hellenic 16: 93–110. Guidoboni, E. and Comastri, A. (1997), The large earthquake of 8 August 1303 in Crete: seismic scenario and tsunami in the Mediterranean area. Journal of Seismology 1: 55–72. Comastri, A., and Traina, G. (1994), Catalogue of Ancient Earthquakes in the Mediterranean Area up to the 10th Century. Istituto Nazionale di Geofisica, Rome. Gutenberg, B. and Richter, C. (1944), Frequency of earthquakes in California. Bulletin of the Seismological Society of America 34: 185–8. Heiken, G. and McCoy Jr., F. (1984), Caldera development during the Minoan eruption, Thira, Cyclades, Greece. Journal of Geophysical Research 89: 8441–62. Imamura, A. (1937), TheoreticaI and Applied Seismology. Maruzen, Tokyo. IOC (Intergovernmental Oceanographic Commission) (1998), Post-tsunami Survey Field Guide. 1st edn. Manuals and Guides 37. UNESCO. Kanamori, H. (1972), Mechanism of tsunami earthquakes. Physics of Earth and Planetary Interiors 6: 346–59. Kastens, K. A. and Cita, M. B. (1981), Tsunami-induced transport in the abyssal Mediterranean Bulletin of the Geological Society of America 92: 845–57. Keller, J. (1978), Prehistoric pumice tephra on Aegean islands, in C. Doumas (ed.), Thera and the Aegean World II. Thera Foundation, London, ii. 49–56. Koutitas, C. G. and Papadopoulos, G. A. (1998), Numerical simulation of the aseismically induced tsunami of 7 February 1963

Tsunamis in the Western Corinthos Bay. Proceedings of the International Conference on Tsunamis, Paris, 26–8 May, 247–54. Lorito, S., Tiberti, M. M., Basili, R., Piatanesi, A., and Valensise. G. (2008), Earthquake-generated tsunamis in the Mediterranean Sea: Scenarios of potential threats to Southern Italy. Journal of Geophysical Research 113, B01301, doi: 10.1029/2007JB004943. McCoy, F. W. and Heiken, G. (2000), Tsunami generated by the Late Bronze Age eruption of Thera (Santorini), Greece. Pure and Applied Geophysics 157: 1227–56. Marinatos, S. (1939), The volcanic destruction of Minoan Crete. Antiquity 13: 425–39. Marinos, G. and Melidonis, N. (1971), On the strength of seaquakes (tsunamis) during the prehistoric eruptions in Santorini, in Proceedings of the 1st International Scientific Congress on the Volcano of Thera, Athens, 15–23 September 1969, 277–88. Minoura, K., Imamura, F., Kuran, U., Nakamura, T., Papadopoulos, G. A., Takahashi, T., and Yalçiner, A. C. (2000), Discovery of Minoan tsunami deposits. Geology 28: 59–62. Monserrat, S., Ibbetson, A., and Thorpe, A. J. (1991), Atmospheric gravity waves and the ‘Rissage’ phenomenon, Quarterly Journal of the Royal Meteorological Society 117: 553–70. Murty, T. S. (1977), Seismic Sea Waves–Tsunamis. Bulletin of the Fisheries Research Board Canada 198: 1–337. and Loomis, H. G. (1980), A new objective tsunami magnitude scale. Marine Geodesy 4: 267–82. Neri, G., Barberi, G., Oliva, G., and Orecchio, B. (2004), Tectonic stress and seismogenic faulting in the area of the 1908 Messina earthquake, south Italy. Geophysical Research Letters 31: 1–5. Papadopoulos, G. A. (1993a), On some exceptional seismic (?) sea-waves in the Greek Archipelago. Science of Tsunami Hazard, 11, 25–3. (1993b), Seismic faulting and non-seismic tsunami generation in Greece. Proceedings of the IUGG/IOC International Tsunami Symposium, Wakayama, Japan 23–7 August, 1993, 115–22. (1998), A reconstruction of the 373 B.C. large earthquake in the western Corinthos Gulf. Proceedings of the 2nd International Conference on Ancient Eliki, Aeghion, Athens, 1–3 Dec. 1995, 479–94. (2001), Tsunamis in the East Mediterranean: A catalogue for the area of Greece and adjacent seas. Proceedings of the Workshop on Tsunami Risk Assessment Beyond 2000: Theory, Practice, Plans. Moscow, 14–16 June 2000, 34–42. (2003a), Quantification of tsunamis: a review, in A. C. Yalçiner et al. (eds.), Submarine Landslides and Tsunamis. Kluwer, Dordrecht, 285–91. (2003b), Tsunami hazard in the Eastern Mediterranean: Strong earthquakes and tsunamis in the Corinth Gulf, Central Greece. Natural Hazards 29: 437–64. (2003c), A tsunami warning system in the SW Aegean Sea, Greece, in J. Zschau and A. N. Küppers (eds.), Early Warning Systems for Natural Disaster Reduction. Springer, New York, 549–52. and Chalkis, B. G. (1984), Tsunamis observed in Greece and the surrounding area from antiquity up to present times. Marine Geology 56: 309–17. and Dermentzopoulos, Th. (1998), A tsunami risk management pilot study in Heraklion, Crete Isl. Greece. Natural Hazards 18: 91–118.

511

and Imamura, F. (2001), A proposal for a new tsunami intensity scale. Proceedings of the International Tsunami Symposium 2001, Seattle, 7–10 August, 569–77. and Kortekaas, S. (2003), Characterisics of landslide gererated tsunamis from observational data, in L. Locat and J. Mienert (eds.), Submarine Mass Movements and their Consequences. Kluwer, Dordrecht, 367–74. and Pavlides, S. B. (1992), The large 1956 earthquake in the South Aegean: Macroseismic field configuration, faulting, and neotectonics of Amorgos island. Earth and Planetary Science Letters 113: 383–96. and Vassilopoulou, A. (2001), Historical and archaeological evidence of earthquakes and tsunamis felt in the Kythira strait, Greece, in G. T. Hebenstreit (ed.), Tsunami Research at the End of a Critical Decade. Kluwer, Dordrecht, 119–38. and Plessa, A. (2000), A new catalogue of historical earthquakes and tsunamis in the Corinth rift, Central Greece: 480 B.C.–A.D. 1910, in G. A. Papadopoulos (ed.), Historical Earthquakes and Tsunamis in the Corinth Rift, Central Greece. Publication 12, Institute of Geodynamics, National Observatory of Athens, 9–119. Imamura, I., Minoura, K., Takahashi, T., Kuran, U., and Yalciner, A. (2004), Strong earthquakes and tsunamis in the East Hellenic Arc. Geophysical Research Abstracts 6: 3212. Karakatsanis, S., Fokaefs, A., Orfanogiannaki, K., Daskalaki, E., and Diakogianni, G. (2005), The 9 July 1956 large tsunami in the south Aegean Sea: compilation of a data basis and re-evaluation, in G. A. Papadopoulos and K. Satake (eds.), Proceedings of the 22nd International Tsunami Symposium, Chania, 27–9 June, 173–80. Papadopoulos, G. A., Daskalaki, E., and Fokaefs, A. (2007a), Tsunamis generated by coastal and submarine landslides in the Mediterranean Sea, in V. Lykousis, D. Sakellariou, and J. Locat (eds.), Submarine Mass Movements and their Consequences. Advances in Natural and Technological Hazards Research 27. Springer, Dordrecht, 415–22. and Giraleas, N. (2007b), Tsunami hazards in the Eastern Mediterranean: strong earthquakes and tsunamis in the East Hellenic Arc and Trench system. Natural Hazards and Earth System Sciences 7: 57–64. Papaioannou, I., Papadopoulos, G. A., and Pavlides, S. (2004), The earthquake of 426B.C. in N. Evoikos Gulf revisited: amalgamation of two different strong earthquake events? Bulletin of the Geological Society of Greece 36: 1477–81. Papathoma, M., Dominey-Howes, D., Zong, Y., and Smith, D. (2003), Assessing tsunami vulnerability, an example from Herakleio, Crete. Natural Hazards and Earth System Sciences 3: 1–13. Papazachos, B. C., Koutitas, C. H., Hatzidimitriou, M. P., Karacostas, G. B., and Papaioannou, A. Ch. (1985), Source and short-distance propagation of the July 9, 1956 southern Aegean tsunami. Marine Geology 65: 343–51. Pararas-Carayannis, G. (1992), The tsunami generated from the eruption of the volcano of Santorini in the Bronge Age. Natural Hazards 5: 115–23. Patacca, E. and Scandone, P. (2004), The 1627 Gargano earthquake (Southern Italy): Identification and characterization of the causative fault. Journal of Seismology 8: 259–73. Perissoratis, C. and Papadopoulos, G. A. (1999), Sediment instability and slumping in the southern Aegean Sea and the case history of the 1956 tsunami. Marine Geology 161: 287–305.

512

Gerassimos Papadopoulos

Pirazzoli, P. A., Ausseil-Badie, J., Giresse, P., Hadjidaki, E., and Arnold, M. (1992), Historical environmental changes at Phalasarna harbour, west Crete. Geoarchaeology 7: 371–92. Piscini, A., Maramai, A., and Tinti, S. (1998), Pilot local tsunami warning system in Augusta, Eastern Sicily, Italy. International Conferernce on Tsunamis, Paris, May 26–8 1998, 137–48. Rothaus, R. M., Reinhardt, E., and Noller, J. (2004), Regional considerations of coastline change, tsunami damage and recovery along the southern coast of the bay of Izmit—the Kocaeli (Turkey) earthquake of 17 August 1999. Natural Hazards 31: 233–52. Shaw, B., Ambraseys, N. N., England, P. C., Floyd, M. A., Gorman, G. J., Higham, T. F. G., Jackson, J. A., Nocquet, J. M., Pain, C. C., and Piggott, M. D. (2008), Eastern Mediterranean tectonics and tsunami hazard inferred from the AD365 earthquake. Nature Geoscience 1: 268–76. Shuto, N. (1993), Tsunami intensity and disasters, in S. Tinti (ed.), Tsunamis in the World. Kluwer, Dordrecht, 197–216. Sigurdsson, H., Carey, S., and Devine, J. D. (1990), Assessment of mass, dynamics and environmental effects of the Minoan eruption of Santorini volcano, in: D. A. Hardy et al. (eds.), Thera and the Aegean World III. Thera Foundation, London, ii. 100–12. Soloviev, S. L. (1970), Recurrence of tsunamis in the Pacific, in W. M. Adams (ed.), Tsunamis in the Pacific Ocean. East–West Center Press, Honolulu, 149–63. Solovieva, O., Go, C., Kim, K., and Shchetnikov, A. (2000), Tsunamis in the Mediterranean Sea 2000 B.C. to 2000 A.D. Kluwer, Dordrecht. Soren, D. (1988), The day the world ended at Kourion— Reconstructing an ancient earthquake. National Geographic, July, 30–53. Thommeret, Y., Thommeret, J., Laborel, J., Montaggioni, L. F., and Pirazzoli, P. A. (1981), Late Holocene shoreline changes

and seismo-tectonic displacements in western Crete (Greece). Zeitschrift für Geomorphologie, Suppl. 40: 127–49. Tinti, S. and Guidoboni, E. (1988), Revision of the tsunamis occurred in 1783 in Calabria and Sicily (Italy). Science of Tsunami Hazards 6: 17–22. and Maramai, A. (1996), Catalogue of tsunamis generated in Italy and in Côte d’Azur, France: a step towards a unified catalogue of tsunamis in Europe. Annali di Geofisica 39: 1253–99. Maramai, A., and Favali, P. (1995), The Gargano promontory: An important Italian seismogenic-tsunamigenic area. Marine Geology 122: 227–41. (2001), Contribution of tsunami data analysis to constrain the seismic source: the case of the 1693 eastern Sicily earthquake. Journal of Seismology 5: 41–61. and Graziani, L. (2004), The new catalogue of Italian tsunamis. Natural Hazards 33: 439–65. Valensise, G. and Pantosti, D. (1992), A 125 kyr-long geological record of seismic source repeatability: the Messina Straits (southern Italy) and the 1908 earthquake (Ms 71/2). Terra Nova 4: 472–83. Van Dorn, W. G. (1968), Tsunamis. Contemporary Physics 9: 145–64. Whelan, F. and Kelletat, D. (2002), Geomorphic evidence and relative and absolute dating results for tsunami events on Cyprus. Science Tsunami Hazards 20: 3–18. Yalçiner, A. C., Kuran, U., Akyarli, A., and Imamura, F. (1993), An investigation on the propagation of tsunamis in the Aegean sea by mathematical modeling, in Proceedings of the IUGG/IOC International Tsunami Symposium, Wakayama, 23–7 August, 65–75. Alpar, B., Altinok, Y., Özbay, I., and Imamura, F. (2002), Tsunamis in the Sea of Marmara—Historical documents for the past, models for the future. Marine Geology 190: 445–63.

This chapter should be cited as follows Papadopoulos, G. A. (2009), Tsunamis, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 493–512.

18

Storms and Floods María del Carmen Llasat

Introduction Floods are the most common natural hazard in the Mediterranean region and they result in the greatest economic losses. For regions such as eastern Spain, the south of France, Italy, and the west of the Balkan Peninsula, it can be argued that damaging floods are not infrequent and they can be considered a key component of the climatic and hydrological regime (Barnolas and Llasat 2007). The economic and social impacts of flooding are not just a function of the high frequency of large floods, but also of the vulnerability created by various human activities. In common with other parts of the Mediterranean, the regions listed above are characterized by widespread and intensive economic activity as well as high population densities—often in river valleys and on coastal plains—and this combination results in very significant losses following major flood events (Ruin et al. 2008). Floods and severe weather events affect all parts of the Mediterranean, but their frequency and impact are not homogeneous across the region. Between 1991 and 1995 the damage caused by floods in Mediterranean Europe amounted to €80 billion (Estrela et al. 2000). Table 18.1 shows the main flood events that have occurred in the European Mediterranean since 1990. Although high-magnitude wind storms are not as important as heavy rains and floods, they are considered to be the second most important meteorological hazard in the Mediterranean region because of their frequency and the damage caused both inland and to sea traffic. In addition to the wind storms produced by well-known winds such as the bora or sirocco, severe-weather events and storms produced by deep lows can affect the northern and southern parts of the Mediterranean basin. The most catastrophic wind storm in recent years was the November 2001 event that caused over 600 deaths in

Algeria and four deaths—and damage put at over €37 million—in the Balearic Islands of Spain. The flooding and wind storms that occur in the Mediterranean and adjacent countries are intrinsically related to both Mediterranean meteorology itself and the marked cyclogenesis recorded in the region (Chapter 3). This chapter begins with a presentation of the forcing factors and, in particular, the role that the region’s physical geography plays in generating such adverse weather phenomena. The section devoted to heavy rainfall events takes account of recent advances in the classification of the convective systems associated with them, based on meteorological radar information (Steiner et al. 1995; Rigo and Llasat 2004; Barnolas et al. 2008). Certain climatic features are included as well as the main meteorological configurations that lead to heavy rainfall episodes, with reference to their synoptic, thermodynamic, and mesoscalar aspects. Detailed discussion of fluvial floods begins with a classification of the types of flooding found in the Mediterranean zone and a presentation of the climatology in relation to both the spatial and temporal distribution of floods. The climatic aspects are linked to a discussion of the long-term record of flooding in the Mediterranean region based on research into historical climatology. Following a section devoted to wind storms, the chapter closes with a short section on tornadoes and hailstorm events in Mediterranean countries.

Cyclones and Meteorological Hazards in the Mediterranean Region It can be argued that the Mediterranean region is geographically and meteorologically one of the most clearly defined regions in the world, with a physical framework that creates a specific Mediterranean air mass and a distinctive Mediterranean meteorology (Jansà 1966,

514

María del Carmen Llasat TABLE 18.1. Major flood events in the European Mediterranean since 1990 Date 1 October 1990 15 November 1990 2 June 1992 22–3 September 1992 26–8 September 1992 3–6 October 1992 31 October 1992 November 1992 23 September 1993 1 November 1993 January 1994 October 1994 10 October 1994 4–6 November 1994 1994 19–25 April 1995 11 August 1995 19 September 1995 4–6 October 1995 19 June 1996 8 July 1996 7 July 1996 7–10 October 1996 8 October 1996 14 October 1996 1997 1998 3–5 May 1999 12–13 November 1999 10–14 June 2000 September 2000 13–16 October 2000 19–25 October 2000 6 and 23 November 2000 9–15 November 2001 8–9 September 2002 29–1 December 2002

Location Southern France Crotone, Italy South-east France Vaisson-la-Romaine, France; Savona, Italy Languedoc-Rousillon, France; Genoa, Italy Veneto, Italy Costa del Tirreno, Sicily Crotone, Italy Liguria, Italy Corcega, France Francia Athens, Greece Catalonia, Spain Piemonte, Italy Greece Piemonte, Italy La Ciotat, France Friuli, France Nîmes, France; Liguria, Italy Versilia, Italy Piemonte, Italy Biescas, Spain Piemonte, Italy Emilia-Romagna, Calabria, Italy Crotone, Italy Badajoz, Spain Sarno, Italy Piemonte, Italy Aude, France; Sardinia, Italy; Catalonia, Spain Catalonia, Spain; Piemonte, Italy Soverato, Italy Piemonte, Val d’Aosta, Liguria, Italy Catalonia, Murcia, Spain Liguria, Italy Algeria Balearic Islands, Spain Gard and other departments, France Lombardia, Italy

1997). The region is surrounded by an almost continuous barrier of mountains that hinder air mass movements. Its position between the cold Euro-Asiatic lands and the warm lands of Africa allows extra-tropical and subtropical influences to alternate and interact (Chapter 3). In addition, the warm surface and deepest water of the Mediterranean Sea favours the interchange of latent and sensible heat (Chapter 2). As a result there are two primary meteorological effects that arise almost continuously (Jansà 1997). First, the formation of a particular low-level Mediterranean air mass, around 1,500–2,000 m in thickness, that is warm and wet from autumn to spring and relatively cold and wet

Number of casualties

Estimated cost of flood damage

42 + 4(dis) 2 4 + 1(dis) 2

336 M€

3

400 MFF 10 M€ 10 M€ 712 M€

14 M€ 64

30 2 1 13

13,000 M€ 14 M€

10 M€

87 1 18 300

5 10

2 >700 5 23 0

65 M€

2,200M dinars >37 M€ 830 M€ 1,200 M€ 13.6 M€

in summer, with the coastlines and the mountain barriers forming almost permanent frontal boundaries. As a result its vertical thermodynamic profile commonly shows potential convective instability. The cooling of the upper layer and/or the lifting of the whole air column can release this latent convective energy and give rise— if the other conditions are present—to heavy rainfall (Llasat and Puigcerver, 1992). Second, the emergence of lee depressions is important as well as the creation of low-level potential vorticity nuclei because of the interaction of the airflow with the mountainous barriers. These can affect the local winds and the production and distribution of rainfall.

Storms and Floods

It is helpful at this point to summarize briefly these meteorological effects from a natural hazards perspective. The Mediterranean region sees the highest concentration of cyclogenesis in the world, the heaviest extra-tropical rainfall events (as much as 800 mm in 24 hours), and very strong local winds with sustained speeds of 20 or even 25 m s−1 (Jansà 1997). In view of these factors, the World Meteorological Organization (WMO) has promoted various programmes on Mediterranean cyclogenesis, such as the Mediterranean Cyclones and Adverse Weather Phenomena Study Project (WMO 1995) or MEDEX—Mediterranean Experiment (Jansà 2002). In most cases, the adverse weather phenomena are connected with both the Mediterranean Sea and the surrounding land and mountain ranges. A good deal of the research into severe weather meteorology and related phenomena in the region has been focused on the areas shown in Figure 18.1a. The area shown in Figure 18.1b will be used to illustrate some examples of high magnitude storms and floods at various points in this chapter. The recent climatology of surface cyclones covering the whole Mediterranean region has been investigated from an operational analysis spanning June 1998 to May 2001 (Gil et al. 2003) for the area lying between 25◦ N and 49◦ N and 12◦ W and 36◦ E (Figure 18.2). This shows that the mean cyclone frequency for the West Zone is about 2,910 cyclones per year, while in the East Zone this figure is 2,248 cyclones per year. If a size filter of 200 km is used in order to eliminate the minor cyclonic perturbations, these figures are 437 and 353 respectively. Figure 18.2 shows the maximum concentrations in the Gulf of Genoa and close to Cyprus. Among other factors, these two maxima are related to the lee cyclogenesis effect produced by the Alps and the Anatolian Peninsula and the coastal mountains of Turkey. Secondary maxima are located south of the Pyrenees, in the Aegean Sea and along the Algerian coast (Figure 18.2). This geographical distribution changes over the course of the year; with summer being the season of greatest frequency and with the maximum concentration of cyclone centres in the south of the Iberian Peninsula and near Cyprus. During the summer, centres of cyclogenesis are mainly observed over land, while in winter most cyclones are located over the sea. The importance of orography in cyclones and cyclogenesis location in Mediterranean areas has been shown by Genovés et al. (1997) based on the application of the HIRLAM (High Resolution Limited Area Model) mesoscale analysis model (0.5◦ network). An initial quantitative statistical assessment of the weight of the orographic factor in Mediterranean cyclone generation shows absolute maxima in the lee of high mountain

515

ranges, where the most intense mesolows are detected. The majority of these are stationary and only a few develop into intense cyclogenesis when a high-level disturbance approaches. A particular case of atmospheric depressions that could be related to some flood events are those associated with cold air pools. According to the first definition by Scherhag (1948), a cold air pool is a deep, cold-core low that is undetectable at ground level. Preferred areas for cold pool genesis are generally associated with either high ground or warm seas (Llasat 1991). These areas shift from the eastern Atlantic in spring to a poorly defined distribution in summer, whereas in autumn and winter the preferred areas shift southwards to the Mediterranean Sea. In this region, preferred areas are to the north-east of the Balearic Islands, the northern Tyrrhenian and Adriatic seas, the centre of the Balkan Peninsula, and the Algerian–Tunisian coast. These regions have recorded more than twelve cold pools per year for a 10-year period (Llasat and Puigcerver 1990). If a cut-off low is defined as a closed depression at altitude that has become isolated and completely detached from the circulation associated with the polar or subtropical stream (and moves independently of that flow), the role that cold pools play in some heavy rainfall episodes are encompassed within the cut-off low definition (in other words, bringing in warm, moist air at low levels) and the vorticity advection. Adverse Weather Phenomena (AWP) are considered to be induced by Mediterranean cyclones when they are recorded at a distance less than 600 km from the cyclone-affected area. Following the WMO (1995) proposal, a meteorological phenomena should be considered an adverse weather phenomena if its intensity, frequency, or extension departs markedly from normal climatic conditions. In the case of high rainfall or wind storms, they become adverse when their intensity exceeds certain thresholds. Thus, it is possible to distinguish between: r Extraordinary: A precipitation input very much

above normal.

r Dangerous: A daily rainfall amount or rainfall dura-

tion period very much above normal. A stormy wind. r Catastrophic: An absolute daily rainfall amount or storm extent very much above normal. Rainfall duration or storm extent very much above normal. The absolute speed of a wind storm above 17.2 ms−1 or covering a wide area. This classification is based on the potential level of destructiveness of the particular adverse weather

516

María del Carmen Llasat

0

250 km FRANCE

Atlantic Ocean

ProvenceAlpesCôte d’Azur

LanguedocRoussillon

Liguria Tuscany Eastern Pyrenees

Catalonia S PA I N

Corsica Levante Sardinia

Andalucia Balearic Islands

(a)

MOROCCO

ALGERIA

ANDORRA

(b)

Adriatic Sea I T A LY Tyrrhenian Sea

TUNISIA

50 km Py

Rajadell C

A

T A

Llobregat

Segre

L O

N

Ter Osor I

A

Barcelona Eb ro Fig. 18.1. (a) The key study areas of adverse weather phenomena discussed in this chapter. (b) The key areas and river systems in north-east Spain and southern France cited in the text.

phenomena (Radinovic 1997). The categories are a function of nature, intensity, and spatial extent. The AWP that are not destructive, but do influence human life and activities are the extraordinary ones. Dangerous weather phenomena are considered to be those

which directly affect human life and material goods. Catastrophic weather phenomena include the dangerous phenomena that affect an extremely large area or reach the absolute maximum of intensity at some station or at some grid point.

Storms and Floods

517

49 21

47 140 97

73

48

64 131 183 74

78

57

114 385 262 227 146 39

53

20

77

30

41

80

75

37

33

88

48

52

21

21

20

23

50

52

40

88

111 61

55

46

67 308 70

40

45 33 38

33

37

43

59

97

24 255 200 151 38 155 81

54

46 162 126 38

81

63 213 82 105 46

26

68

46

84 113 34 226 64

61

86

25

88

35

98

41 32

42

48 157 137 58

21

28

72

24

61 108 98 130 24

38

42

34

70 135 148 109 140 38

34

51

53

78 101 70

22

68

78

37 49

159 252 118 162 48

93 150 238 38

47

24

20

67 188 532 361 450 37

30

20

32 146 96

33 68 432 93

44

23

21

389 207 270 176 61

30

24

30

32

73

32

50

37 34

34

24

41

31

62

29 73

79

58

67

43

37

25 -12

-8

35 107 64 115 49

-4

0

57

28

42

28

24 136 172 120 43

4

8

65 245 189 130 339 438

12

16

20

24

28

32

36

49 5

7

7

6

6

7

5

5

5

9

8

5

5

10

14

6

7

5

6

6

16

9

6

5

19

10

7 8

7

5

7

5

8

6

9

6

8

6

6

7

10

10

7

11

33

6

11

52

11

5

8

45 6 5

16

15

90

30

17

31

25

22 14

63

8

6

34

23

5

22

18

11

17

5

8

9

11

6

7

7

6

6

9

5

14 5

9

41 5

12

8

18

6

6

15

17

17

12

8

8

11

7

13

17

20

16

14

8

29

24

11

13 5

23

21

8

14

11

7

13

13

67

14

53

6

9

54

58

19

26

37 6

8

111 82

16

8

18

20

6

9

7

33 15

6

5

5

6

39

26

13

14

17

8

21

22

10

22

91

44

25

7

9

17

42

34

12

6 5

6

29 39 5

10

22

5

108

25 -12

-8

-4

0

4

8

12

16

20

24

28

32

36

Fig. 18.2. The mean annual frequency of cyclones in the Mediterranean in the summer (upper map) and in the winter (lower map). Cyclone centres have been mapped in a grid of 2◦ latitude by 2◦ longitude for the period June 1998 to May 2001 (modified from Gil et al. 2002).

Jansà et al. (1996) have identified the presence of a cyclonic centre in 79 per cent of 721 heavy rainfall events in the Mediterranean region. Recently, Rigo and Llasat (2003) have shown that 79 per cent of heavy rainfall events occurring in Catalonia, Spain, were related to a cyclone. In the vast majority of cases, Mediterranean depressions form on the coast of Algeria due to the interaction of the wind with the Atlas Mountain chain. In cases where the rains affect Catalonia, Languedoc-Rousillon, or Provence-Alpes-Côte d’Azur (Figure 18.1a), the depression then shifts northwards. In order to lift the mass of warm, moist air there has to

be a forcing mechanism, which can be either meteorological (e.g. a convergence line, front, dry lines, etc.) or topographical (e.g. the coastline or, more usually, the mountains). It has been observed that the position of the depression plays a decisive role when it comes to localizing the zone in which the storms will occur, especially in the case of the southern Mediterranean. The analysis of some recent episodes has shown how the storms can shift with the movement of the depression (Llasat et al. 2003a). Contrary to the common view that is mainly promoted by mass media, the presence of a cold pool or cut-off

518

María del Carmen Llasat

low is not a necessary condition, nor is it sufficient on its own, for the occurrence of heavy rain. When cutoff lows are associated with such events, their role is to create a rather deep cyclonic circulation in the lower troposphere so that Atlantic air may be carried over the Mediterranean Sea. If the upper vortex is in the right location, this air, which is warm and very humid in the lower layers, impinges almost at right angles on the coastal mountain ranges and the forced ascent is enough to trigger the potential instability. Rainfall amounts over 200 mm in 24 hours are not uncommon in such cases, particularly on the eastern Spanish coast, where 40 per cent of catastrophic floods are associated with the presence of a cold pool. Due to the confusion generated by incorrect use of the term ‘cold pool’, some Meteorological Offices (Martín León 2003) have recently suggested using the term DANA (Depresión Aislada en Niveles Altos, Isolated Depression at High Levels) for those that could be identified as a cut-off low or COL (Llasat et al. 2007).

Convective Systems and Heavy Rainfall Events in the Mediterranean Region Beginning with the initial classification made by Byers and Braham (1949), and incorporating the most recent classifications (Doswell et al. 1996; Rigo and Llasat 2004), along with data from meteorological MCS-TS

radar, heavy rainfall events recorded in Mediterranean areas can be associated with the following structures (Figure 18.3): 1. Mesoscale Convective System (MCS): a precipitation structure may be identified as an MCS when its major axis has a length equal to or exceeding 100 km for three hours or more, and a minimum of 30 per cent of its area in each radar image can be associated with convective rainfall. An MCS is classified depending on the position of its stratiform region in relation to the convective area, and depending on the organization of the convective region. Thus, they could be divided into wellorganized systems (usually organized lineally) and poorly-organized systems, or clusters of convective structures (CLU). The first type could be divided into TS (with trailing stratiform area), LS (leading stratiform region), and NS (with practically no stratiform precipitation). A specific case of the NS class would be when the stratiform region is located on a flank of the convective line and its movement is parallel to the movement of the convective region. 2. Multicell systems (MUL): if a minimum of 30 per cent of the area covered by the precipitation structure in each radar image can be associated with convective rainfall, but does not meet the time and size conditions of an MCS.

MCS-LS

MUL

IND

EST- EMB

EST

Convective Stratiform MCS-NS

MCS-CLU

No rainfall

Fig. 18.3. Classification of cloud systems associated with heavy rainfall events in the Mediterranean area (after Rigo and Llasat 2004). Key: MCS (mesoscale convective system), CLU (clusters of convective structures), TS (with trailing stratiform area), LS (with leading stratiform regions), NS (with practically no stratiform precipitation), MUL (multicell systems), IND (isolated convection), EST-EMB (convection embedded within stratiform rainfall), and EST (stratiform).

Storms and Floods

3. Isolated convection (IND): when small-scale, independent, and separate convective structures are identified. 4. Convection embedded in stratiform rainfall (ESTEMB): a stratiform region with some convective nuclei. The area covered by the convective precipitation does not exceed 30 per cent of the total area covered by the precipitation structure. 5. Stratiform (EST): when convective precipitation does not exist or does not exceed 3 per cent of the total area covered by the precipitation structure. 6. Supercell structures: if the precipitation structure has one large convective nucleus and a mesocyclone is detected within it (Doswell and Burgess 1993), a feature only detectable by using Doppler wind information. At present, the only systematic analysis of convective systems and heavy rainfall events made in the Mediterranean region has been done for the north-east of Spain (Rigo and Llasat 2004) (Figure 18.1). This analysis shows a total of 167 main precipitation structures associated with thirty-one heavy rainfall events recorded between 1996 and 2000. MCS and Multicells are the most common structures, with fifty-seven cases in each category, followed by individual cells (thirty-two cases). The cluster structure is the most frequent among the MCS cases (more than 50 per cent). Although it is possible to record all the structures in any season of the year, they arise mainly during the autumn season. An exception has been identified for the stratiform structures, which are not recorded in either spring or autumn when heavy rainfall events are produced. In the same region, an analysis of the longest rainfallrate series in Europe (1927–81) shows that 92 per cent of rainfall episodes in Barcelona (Spain) are nonconvective, yielding 63.5 per cent of the total precipitation, while 8 per cent are convective and provide 36.5 per cent of the total precipitation (Llasat 2001). To break this proportion down further, 3.8 per cent pertain to slightly convective episodes (less than 25 per cent of the total precipitation has a 5-min intensity of 35 mm per h), 2.9 per cent to moderately convective episodes (between 25 per cent and 75 per cent of the total precipitation above the previous threshold), and 1.3 per cent to strongly convective episodes (more than 75 per cent of the total precipitation exceeds 35 mm per hour). The intensity and duration ranges established by the above classification coincide broadly with the proper-

519

ties of the various convective systems. The episodes termed strongly convective, of short duration and high average intensity, would have their origin in unicellular or multicellular storms (the life-cycle of a typical storm cell ranges between 20 and 30 minutes), arising mainly in summer. The moderately convective episodes that are of high mean intensity—with a duration exceeding a few hours—would correspond to highly organized convective systems composed either of multicellular or Mesoscale Convective Systems in which it is possible for 50 per cent of the precipitation to be stratiform.

A General Conceptual Meteorological Model Associated with Heavy Rainfall in the Mediterranean Region Synoptic and mesoscale factors responsible for heavy rainfall are not always the same either within a region or in different regions. It is nevertheless possible to speak of a basic conceptual model associated with those heavy rainfall events that produce catastrophic floods in the north-western Mediterranean region. This conceptual model (Llasat and Puigcerver 1992; Jansà et al. 1995, 1996; Palmieri and Clericci 1992) shows a long anticyclonic situation over the Mediterranean—for the days leading up to the events—which favours the formation of a Mediterranean air mass (Figure 18.4). Normally, the presence of any Mediterranean low or a convergence line organizes the differentiated air currents as well as internal low frontal boundaries. The intersection between the tip of a warm-wet current and a thermal-humidity boundary is the most likely place for reaching or releasing the convective instability and for the development of large convective clouds producing heavy rain. If the situation remains more or less stationary, the accumulated rainfall can reach very large amounts. Another possibility is to substitute the low internal boundary with a mountain barrier that stops and forces the ascent of the warm-wet current. Usually, both variants appear in combination and it is difficult to assess the relative contribution of each factor, bearing in mind the synergetic effects between them. As Figure 18.4 shows, the presence of the anticyclone over the central Europe–north Mediterranean area encourages the entry of a low-level flow from the south or south-east over the region, as well as the evaporation and accumulation of water vapour over the western Mediterranean. The Mediterranean is a semi-enclosed

(a) 500 hPa

-24

-16

6

6

5,880 5,800

-20

-1

5,720

2

5,640

-24

5,560

-12

5,480 5,400

-8

5,320

4 8

(b) 850 hPa

12

Temperature in °K

0

8

4

8

16

0

1,660 1,620 1,580 1,540 1,500

16

1,460

16

1,420

20

1,380

4

1,340

8

12

Geopotential height (m)

5,960

24

Geopotential height (m)

-1

Temperature in °K

-1

-20

1,300

16

12

1020 1020

101

10

(c)

6

Atmospheric pressure (hPa)

10

10

110016 6

1012 12 10

20

1024

16

088 1100

1100 166 101

2

11001166

Fig. 18.4. An example of the synoptic situation that produces heavy rainfall with catastrophic floods in the western Mediterranean region. The maps are for 10 June 2000 at 00.00 UTC (Universal Time Corrected) and show the region at (a) 500 hPa and (b) 850 hPa. Note that 1 hPa (hectopascal) = 1 millibar. Temperature is given in ◦ K in (a) and (b). (c) Mesoscale surface analysis with pressure in hPa. The mesoscale low encountered warm Mediterranean air over the coast at low altitude. The presence of relatively cold air at 500 hPa and warm and wet air between 850 hPa and the ground was conducive to instability. A convergence line forced by the mesoscale low triggered this instability that was reinforced by the coastal mountains.

Storms and Floods

sea of no great depth, so it forms a large source of heat and moisture. Moreover, in many recent storms it has been observed that the temperature of the sea is even higher than usual and it is now well established that heat transfer and thermal convection reinforce cyclogenetic processes in the Mediterranean region (Alonso-Sarría et al. 2002). Typically, there is an Atlantic depression situated in the western part of the Iberian Peninsula or France (with an associated front that moves towards the north-east) and that helps to strengthen this flow. The position of the anticyclone and/or low is essential in organizing the flow that will affect one region or another. Romero et al. (1999) have shown that the majority of torrential rainfall events produced in Mediterranean Spain (Chapter 8) are associated with the presence of closed cyclonic circulations or shortwave troughs located to the south or west of the affected area. Good examples of this type of situation affecting Italy are two different events which impacted the Tanaro River basin in north-west Italy on 14–18 November 2002 and 24–6 November 2002. The synoptic scenario involved a sequence of two deep troughs progressively extending their influence over the Mediterranean basin. In both cases the flow became southerly and moist air impinged against the Alps and Apennines generating heavy precipitation. A further deepening of the second trough associated with a cut-off event took place between 25– 6 November. As a result, a deep depression was isolated to the south of Sardinia and produced wet south-easterly flow from the Tyrrhenian Sea towards the western Alps and Apennines (Taramasso et al. 2005). It is quite common to find moving systems that, in the course of their journey, affect different Mediterranean regions or countries. In Spain, for example, a convective system can start in the south and move towards the north-east, affecting all the regions along its path. This was the case in the flooding of 19 October 1982 which claimed thirty-nine victims in the Levante region of Spain, with over 500 mm of rainfall in 24 hours. This same system had previously produced rainfall exceeding 100 mm in 24 hours in Andalucia and did the same in Catalonia. Similarly, heavy rainfall systems that affect both Catalonia and the south of France are not infrequent (Llasat and Puigcerver 1994). An example was the serious flooding that occurred during 6– 8 November 1982 and the regional pattern of rainfall for that event is shown in Figure 18.5. In addition to affecting the Eastern Pyrenees region (Figure 18.1b), where 610 mm of rainfall were recorded over the three days, 150 mm (more than double the monthly average) were recorded in southern Andalucia on 6 November. A similar situation had arisen in October 1940, with

521

more than 800 mm in 24 hours in the Eastern Pyrenees (Llasat 1993). Similarly, the heavy rainfall events that occurred in October 1986 and 1987, and again in November 1988, with more than 200 mm in 24 hours, caused floods in both the French and Spanish zones of the Eastern Pyrenees (Llasat and Puigcerver 1992; Ramis et al. 1994, 1995). Floods that are caused by a single meteorological event can affect Spain, France, and Italy and this was the case in September 1992. The first heavy rainfall events were recorded in the south of Catalonia and in Levante (eastern Spain) on 26 September, with total quantities exceeding 100 mm in many places. In the Rousillon region of France, the maximum rainfall recorded was 324 mm, 93 mm of that in 1 hour. On 27 September the rains mainly affected the Liguria region of Italy, where a total of 459 mm was recorded in 24 hours, with a maximum mean hourly intensity of 75 mm. On 28 September the rains mainly affected the Tuscany region of Italy, with a maximum of 148 mm. References to the direct role played by orography in heavy rainfall events are frequent. In the case of the 26–8 September 1992 event, the coastal mountains (over which the wind impinged perpendicularly) triggered the convection which remained latent over the Mediterranean Sea, with high CAPE (Convective Available Potential Energy) values, strong quasi-geostrophic vertical forcing, water vapour nourishment, and instability at low and medium altitudes (Llasat et al. 1999). In the Versilia event (Tuscany, Italy), recorded on 19 June 1996, heavy precipitation over the course of twelve hours occurred (with a maximum accumulated value of 478 mm) as a consequence of strong potential instability with large CAPE values and a low-level jet mainly driven by the local topography (Frontero et al. 1997). Sometimes the focusing of the convection and the quasi-stationary nature of the system can be explained by taking into account the relationship between the mountains and the dominant flow at low levels (Riosalido et al. 1997). On those occasions, the slow translation of the system is a consequence of the weak wind and the formation of new cells in the opposite direction to the movement of individual cells, in conjunction with winds blowing perpendicularly to the mountain ranges which are more than 2,000 m above sea level. This was the case for the flash flood that occurred in Biescas, Huesca (Spain) in the Central Pyrenees, on 7 August 1996. The flood completely destroyed a campsite that was situated between two small gullies that overflowed. Out of a total of 630 people registered at the campsite, 183 sustained

María del Carmen Llasat

11880

110000

220 220

460 460

260 260

14 1400

610 610

1000 10

140 140

80 80

11400

1100 114400 00

114400 323 323

334400 26 260 0 2200 22 11800 114400

Girona

2200 100 100

80 80

Aragon

60 60

576 576

600

10 10

118800

00 110

11400

0 22220

282 282

140 14 0

8800

0 1100

110000

22200

11800

110000

412 412

118800 60 60

11800

10 1000

0 1100

226600

220 220

80 80

80 80

22200

118800

522

Lleida

60 60 1100

Catalonia

50 km

Tarragona

60 60

Barcelona

Mediterranean Sea

Fig. 18.5. Rainfall distribution for a flood event of Type 2b that took place from 6–8 November 1982. The map straddles the Spain– France border and shows the Eastern Pyrenees, including Andorra, the southern part of Languedoc-Roussillon, and north-east Spain. The maximum accumulated rainfall of 610 mm in less than 72 hours was recorded in Py in France (Figure 18.1b). Catastrophic floods occurred in rivers of Types A and B. In Spain alone, flood damage amounted to over 270 million Euros and fourteen people died.

injuries and a further 87 died in less than 45 minutes. The maximum precipitation was 269 mm in 24 hours, with 226 mm in only 3 hours and 153 mm between 16 and 17 hours UTC. However, orography can also play an indirect role in heavy rainfall development. Between 28 September and 5 October 1987, 431 mm of rain was recorded in Barcelona (Catalonia, Spain). On that occasion the main role of the Pyrenees was related to the generation of a strong orographic dipole, giving rise to a strong mesohigh over Catalonia. As a consequence of this high, the synoptic south-easterly wind at low levels was substituted near the Catalan coast by a westerly wind and a convergence line developed over the Mediterranean Sea. The situation remained quasi-stationary for a few days and the continuous vertical forcing of the convective and unstable wet air gave rise to the generation of an MCS (Ramis et al. 1994). In contrast, in eastern Mediterranean countries where heavy rainfall events with more than 100 mm in 24

hours are rare phenomena, the role of orographic factors is secondary or non-existent in rainfall production (although it can play an important role in cyclogenesis). For example, an extreme rainfall event (maximum input of 44 mm per day in the Balkans and Asia Minor) over the Black Sea, Ukraine, and Russia occurred from March 28 to 1 April 1995. The event started in the Mediterranean, when a secondary cyclonic eddy was formed in a deep trough over northern Italy and moved over the Adriatic Sea, although the main factor was a frontal system driving the circulation (Chakina and Berkovich 1997). In Greece, many floods are produced by heavy but short rainfall events, and the role of deforestation and urbanization is often very important in their genesis. There is some evidence to suggest that they mainly affect urban areas, due to the transformation of urban catchments and river channels. They are more destructive in the western parts of Greece due to climatic, geomorphological, and vegetation factors and the impact of human

Storms and Floods

activity. The most serious flood of the last ten years in Greece was in 1994, when a thunderstorm produced 68 mm of rain in one hour in Athens. Unlike the western part of the basin, in the eastern part of the Mediterranean, the main flood events commonly occur in spring.

Classification of Floods and Storms in the Mediterranean Region Floods are a complex hydrometeorological hazard. Meteorological and climatic factors, drainage basin factors, drainage network and channel morphometrics, and human factors play a major role (López-Bermúdez et al. 2002). Heavy rains, long rainy periods, or snowmelt are necessary but not sufficient to cause them. Other factors such as antecedent precipitation, terrain, and surface run-off characteristics are also important (Chapters 6 and 8). Natural processes also interact with human activities. For example, land use and its history and the civil and hydraulic infrastructure can have very variable effects on the natural patterns of flood generation and impact. It is not easy, therefore, to create a Mediterranean flood classification that can be applied to all catchments and events. The classification presented here attempts to integrate some of the meteorological, hydrological, and impact aspects of Mediterranean floods, although hydraulic aspects have not been considered. In order to understand the various types of flood events that can affect Mediterranean countries, it is important to remember from the outset that Mediterranean river systems comprise three basic types: r Type A: High-mountain rivers (>2,000 m) with

rainfall- and snowmelt-influenced regimes with large basins (>2,000 km2 ) and perennial flows. r Type B: Rivers with upland headwaters with intermediate-sized catchments (50–2,000 km2 ) and flows. Due to the highly seasonal rainfall regime (with an extended dry season) and the abstraction of river flows, some have ephemeral reaches in their catchments. r Type C: Short littoral water courses with small catchments (5–50 km2 ) and steep channels. They are ephemeral systems marked by infrequent but torrential flows. A key control on the behaviour of these rivers is the torrential rainfall events discussed in the first part of this chapter and inherent to the Mediterranean region. Accordingly, a distinction can be made between four kinds of floods and associated flood-generating mech-

523

anisms in Mediterranean catchments, and these are described in turn below.

Type 1: Short-lived, High-intensity Rainfall Events The first type of storm involves short-lived events (less than 3 hours and often less than 1 hour) of very intense precipitation (peaks of rainfall intensity above 3 mm per minute) but limited overall rainfall totals (<100 mm). This kind of event requires considerable local instability. They are ‘strongly convective events’ produced by ‘isolated cells’ or ‘multicells’ of limited horizontal extension. Despite limited extension, they sometimes develop into an unstable atmospheric environment that can produce rainfall, strong wind, or hail in other locations of the surrounding region. As they are not widespread, following the Radinovic classification, they could generally be considered as ‘dangerous’. They appear during summer and early autumn and produce ‘local flash-floods’ predominantly in rivers of Type C. In these events the peak flow is more or less simultaneous with the rainfall peak. Although total discharge may not be very high, flood height (stage) can be considerable due to the narrow river channels, steep catchments, and, in some catchments, urban structures and street layouts. These events mainly affect the north-west part of the Mediterranean (Figure 18.1) in densely inhabited and tourist regions. Even though in the western part of this area floods are less frequent, local flash floods of this type have a major impact, due to the intensive urbanization of some areas in which seasonally dry or ephemeral channels are found. These events can bring road traffic to a standstill, give rise to power cuts (such events are usually combined with thunderstorms), and sweep away cars parked in the littoral water courses or in adjoining streets. Any loss of life is usually the result of the reckless behaviour of people who underestimate the power of flood waters and attempt to cross the water course. This kind of event can also affect small mountain catchments. The flood that affected Athens on October 1994, with damages in excess of €14,000,000, was of this type. Figure 18.6 shows a heavy rainfall event that produced a sudden and very local flash flood on 3 April 1989 in a basin of less than 100 km2 in the headwaters of the Llobregat River in the Eastern Pyrenees.

Type 2: Sustained Heavy Rainfall Events Type 2 events are related to episodes of heavy rain sustained for several hours that can produce catastrophic

(a) 2000 C

8 1500

5

1500

7

2000 B

9 S

1000

P

6 10

11

0

150

1

2 3

1000

V

0

150 00 10

4

2000

N

E

0

5 km

Llobregat River

(b)

25

6

30 2

40

4 50

20 15

Intensity (mm/min)

(c) 6 Vallcebre (2)

Carot (4)

Vilella (6)

4 2 0 0

6

12

18

24

0

6

12

18

24

0

6

12

18

24

Hour Fig. 18.6. An example of a heavy rainfall event that produced a flash flood event of Type 1 (3 April 1989) in a mountain catchment in the headwaters of the Llobregat River on the southern margin of the Pyrenees. The monitoring network shown is part of the Vallcebre catchments programme (Gallart et al. 2002). (a) Map of the topography and drainage network and the monitoring stations. The rain gauges are numbered 1–11. The letters mark the location of small villages or country houses. C and V also have rain gauges. (b) The movement of the storm across the basin. (c) Rainfall intensity data from rain gauges 2, 4, and 6. Their locations are shown on maps (a) and (b).

Storms and Floods

floods due to the higher than normal daily rainfall amount (usually more than 200 mm) and the extensive area covered by the storm (usually >2000 km2 ). This kind of event requires convective instability with abundant feeding of warm and wet air from low levels, and a mechanism to force air ascent to release the potential instability or to destabilize the air column. Convective rainfall is generally produced by ‘multicells’ or ‘mesoscale convective systems’. It is possible to distinguish between two kinds of flood event in this category. Type 2a lasts less than 24 hours and the maximum precipitation is usually recorded in less than 6 hours, with accumulated rainfall of nearly 200 mm. They are ‘strongly convective events’ and can produce catastrophic flash floods in Type B and Type C rivers that are simultaneous with the maximum rainfall input. Type 2b events last more than 24 hours but generally less than four days. Although accumulated rainfall usually has values between 200 and 400 mm, values of more than 800 mm are possible. Peaks of strong rainfall intensity and moderate but continuous rainfall are recorded successively. Consequently, they are ‘moderate convective events’, and while they can also produce local flash floods in Type C rivers the floods occurring in Type B river catchments (and occasionally Type A rivers also) are the most significant. Floods of Type 2a have produced the highest number of casualties when they have affected flood-prone areas with high concentrations of people. Damage produced by catastrophic floods of Type 2b relates to total or partial destruction of infrastructure (e.g. houses, bridges, and roads), power cuts, urban inundation, agricultural and livestock losses, and, frequently, loss of human life. These events usually occur in autumn, although some cases have also been recorded in spring or summer. Figure 18.7 shows the rainfall map and the changes in rainfall intensity associated with the catastrophic flood that affected Catalonia on 10 June 2000. The maximum cumulative precipitation was 224 mm and this produced a peak flow of 628 m3 s−1 in the ephemeral channel of the Riera de la Magarola (a 96 km2 tributary basin of the Llobregat River) that caused extensive damage (Figure 18.8). The peak discharge of the main channel of the Llobregat River at its mouth was estimated at 1300 m3 s−1 . The maximum recorded peak flow for the Llobregat River of 3080 m3 s−1 was reached during the Type 2b floods of 20–3 September 1971. Another example of a Type 2a event was the storm of October 1988 that produced a flash flood on the Cadereaux River in southern France. On 3 October the flood peak reached 600 m3 s−1 in a catchment area of only 42 km2 . The flood destroyed many cars and some buildings in the city

525

of Nîmes in southern France. The event was produced by a rainfall input of 300 mm in six hours (Stanescu, personal communication). The flash flooding produced in southern France on 8 September 2002, with 680 mm of rain in 24 hours, was also related to a Type 2a event (Milelli et al. 2006). During this event an area of 5,393 km2 recorded more than 200 mm of rainfall and large areas were inundated across the Gard, Herault, and Vaucluse departments in southern France (Ruin et al. 2008). Figure 18.9a shows extensive flooding in the village of Aaramon near Nîmes and Figure 18.9b shows flood waters of the River Crieulon over-topping a dam near Sommières. In Liguria in north-west Italy on 13 August 1935, a precipitation event of 389 mm in eight hours produced flash flooding along the Orba River (141 km2 ) with a peak flow of 2200 m3 s−1 . This is also a clear example of Type 2a event. The floods recorded during 14–18 and 24–6 November 2002, with maximum accumulated rainfall in 72 hours of 560 mm and 442 mm respectively, could also be considered as Type 2b events. During September 1992, two catastrophic Type 2b flood events occurred over the south of France and the north of Italy. The first took place between 22 and 23 September and the peak flow of the flash flood produced on the Ouvèze River (580 km2 ) in Vaisson-la-Romaine (PACA), was above 1000 m3 s−1 . The second event started on 26 and ended on 28 September in central Italy. Another catastrophic flood event of Type 2b was recorded during 12–13 November 1999 in the Aude region (LanguedocRoussillon), with 624 mm in 36 hours.

Type 3: Long Duration, Low-intensity Rainfall These are episodes of long duration (approximately one week) with relatively weak average rainfall intensity values, although there may be peaks of high intensity within the overall distribution. Total precipitation during Type 3 events can be >200 mm. If floods occur, they are usually in larger rivers of Type A and B as described above. They may be described as ‘slight convective events’ and are usually associated with convection embedded in stratiform rainfall (Figure 18.3). Although not very frequent, they usually occur in winter and, occasionally, in spring. The floods recorded in Catalonia in January 1996 (Llasat et al. 2000) are a good example (Figure 18.10). River catchments on the Balkan Peninsula have also seen Type 3 events. The city of Celje in Slovenia, for example, that lies on the confluence between the Savinja (1192 km2 ) and Voglarinja

526

María del Carmen Llasat

(a)

Precipitation (mm) 200 180 160 140 120 100 80 60 40 20 0

50 km

(b)

5 minute intensity (mm/hour)

125

10 June 2000

100

75

50

25

0 0100

0300

0500

0700

0900

1100

1300

1500

Time Fig. 18.7. A flood event of Type 2a that took place (10 June 2000) in Catalonia. (a) Rainfall distribution over Catalonia during the event. (b) Rainfall intensity during the course of the event recorded at the Rajadell station in the Llobregat basin (Figure 18.1b). A depth of 224 mm of rain fell in <12 hours. Catastrophic flooding occurred in several rivers, with five fatalities and damage amounting to over 65 million Euros.

Storms and Floods

Fig. 18.8. Destruction of a road bridge by the Riera de la Magarola (an ephemeral tributary of the Llobregat River) during the flood event, 10 June 2000 (photo: J. Guamis).

(413 km2 ) rivers, was struck by a large flood in June 1954 after three weeks of rainfall, when most of the city was inundated (Brilly and Polic 2005).

Type 4: Snowmelt Events Flood events produced mainly by melting snow are most common in Type A river basins. It should be noted that the melting of snow in the mountainous headwaters of river basins produces regular and predictable increases in flow only during the spring, and only occasionally gives rise to overbank flows when accompanied by heavy rainfall (rain on snow events) or moderate but continuous rainfall. Their contribution is particularly important in the central and east-

527

ern part of the Mediterranean basin. An example would be the floods and landslides recorded in the Lombardia region of Italy following the rainfall and snowmelt produced between 14 November and 1 December 2002 (Fossati 2004). Other kinds of floods not considered in the above classification are those produced by rapidly rising groundwater levels in karst terrains (Chapter 10), dam bursts (Chapter 6), and coastal flooding from storm surges (Chapter 13) and tsunamis (Chapter 17). The types of rain responsible for the majority of floods in the Mediterranean region are those of great intensity that take place over a short period of time over small catchment areas with steep slopes (Type 1 floods). In some catchments, snowmelt can increase the duration and peak of the flood. This type of flood, most typical in the northern Mediterranean and central Europe, is also important in the lowest reaches or near the mouths of rivers such as the Ebro, Rhône, or Po. Figure 18.11 shows the relationship between extraordinary floods and average yearly floods for basins of different sizes situated in the Mediterranean and non-Mediterranean regions of Europe. In general, in Mediterranean catchments, there is a much greater difference between the highest recorded flows and the average annual flows for catchment areas <10,000 km2 .

Long-term Studies and the Impact of Climatic Variability on Floods The Intergovernmental Panel on Climate Change (IPCC 2001) report shows that an increase in climate variability and some extreme events could

Fig. 18.9. Some consequences of the flash flooding produced in the Gard region of southern France, 8 September 2002. (a) A flooded village near Aaramon, and (b) the floodwaters of the River Crieulon overtopping the Sommière dam (photos: courtesy of the Centre Méditerranéen de l’Environment).

528

María del Carmen Llasat

(a)

Precipitation (mm) 250 200 150 100 75 50 25 0

50 km

(b)

5 minute intensity (mm/hour)

30

20

10

0 21

22

23

24

25

27

28

29

30

January 1996 Fig. 18.10. A flood event of Type 3 January 1996 (21–30) recorded in Catalonia, north-east Spain (Figure 18.1a). (a) Map of rainfall input during the event and (b) the temporal pattern of rainfall intensity over the nine-day period recorded at the Boadella reservoir station in the Muga River valley. During this storm 237 mm of rainfall were recorded at Osor in the Ter basin (Figure 18.1b).

Storms and Floods

Relationship between maximum flow and average flow

1,000

100

10 Mediterranean Non-Mediterranean

1 100

1,000

10,000 100,000 Basin area (km2)

1,000,000

Fig. 18.11. The relationship between maximum flow and average annual catchment flow for river catchments of various sizes in the Mediterranean and non-Mediterranean regions of Europe (modified from Estrela et al. 2000).

be one consequence of the increasing atmospheric concentrations of greenhouse gases and aerosols. Thus, new scenarios of increased climatic variability and global change justify a detailed reconstruction of past flood events from proxy-data. It is now widely appreciated that the modern instrumental period cannot adequately cover natural phenomena characterized by a very low frequency of occurrence. Although some of the earliest research in Europe into past flooding based on historical documentary sources took place in the second half of the nineteenth century, it is only relatively recently that historical floods have been considered in the Mediterranean region alongside a modern understanding of climate dynamics and meteorology (Pavese et al. 1992; Camuffo and Enzi 1996; Lang and Cœur 2002; Barriendos and Martin-Vide 1998). The flood chronologies that can be constructed from documentary records for the Mediterranean (under optimum conditions) do not usually extend further back than the fourteenth century, except for those in Italy dating from the Roman Empire. However, other sources, such as those obtained from landform- and sedimentbased palaeoflood analysis, allow the period of study to be extended back some thousands of years into the Holocene period (Benito et al. 2003; Thorndycraft et al. 2003; Benito and Thorndycraft 2005; Maas and Macklin 2002; Chapter 11). The major documentary sources containing climatic information and details of its effects are local and state government records, religious collections, private collections, and notaries’ archives (Barriendos et al. 2003). Documentary series from municipal authorities contain

529

most of the available information. However, strict criteria must be applied to obtain the best data series of the highest possible quality. The information collected usually includes data on the meteorological phenomenon that was the main cause of the overbank flows, the exact dates and duration of the flood event, and detailed descriptions of the direct and subsequent impacts on infrastructure and population. Quantitative references to the behaviour of the rising waters are sometimes found, such as the levels reached or the zones flooded. In this case, it is important to bear in mind that flood levels are very sensitive to hydraulic works or channel-bed incision and deposition, which might have changed the relationship between water level and discharge. Also, flood damage is produced by a combination of flood hazard and land use activities. It is also important to appreciate that an increase in flood vulnerability, due, for example, to urbanization in the flood-prone area, will lead to more frequent damage even if the natural flood hazard remains stable. Whenever possible, the historical flood classification should be based on discharge estimates, with a sensitivity analysis to assess the specific errors of the hydraulic model for the conversion of historical flood levels into discharge. However, because the general topographic map coverage of the main countries in Europe dates from the eighteenth century, with reliable height information from only the second half of the nineteenth century, in many cases it is difficult to gain a precise knowledge of the river channel topography over several centuries. Furthermore, some specific problems related to the availability of old data for model development and calibration, with uncertainties on the expected differences between the past and the present configuration of the river and catchment areas, may hinder discharge estimation. Thus, in order to have the longest possible flood series, a scale of event magnitude can be proposed using the effects of the floods on the river channel system and surrounding areas. This is the approach used in the following classification. It comprises three levels based upon hydrological criteria and the impact of the flood (Barriendos and Llasat 2003b). 1. Ordinary rise in stage or small flood: The rise in stage does not usually involve the river channel overflowing. No serious damage or destruction is caused, but minor damage to hydraulic installations such as mills or irrigation channels cannot be ruled out. 2. Extraordinary to intermediate flooding: The discharge is sufficient to exceed channel capacity. Floodwater is present in urban areas or in the

530

María del Carmen Llasat

reaches under study, although the level reached in them can be variable. Minor damage to urban infrastructure, but the impact on hydraulic installations adjacent to or in the channel, such as mills, irrigation channels, dams, or footbridges can be severe, with partial destruction. 3. Catastrophic flooding or large flood: The levels reached can be the same as, or exceed those of, an extraordinary flooding episode. The difference lies in the strength or capacity of the out of channel flows to cause severe damage or complete destruction of infrastructure close to the river. Furthermore, the floodwaters can affect zones away from the channels and with destructive effects. Dwellings may collapse, sections of roads may be destroyed, crops may be lost, and shrub and tree species may be uprooted and swept away. This classification has been applied to three major rivers in Catalonia: the Segre (22,579 km2 ), which is a tributary basin of the Ebro River, the Llobregat (4,948 km2 ), and the Ter (3,010 km2 ) (Llasat et al. 2005). It has also been applied to two rivers of the south-east of France: the Isère (5,700 km2 ) and the Ardèche (2,372 km2 ) (Barriendos et al. 2003). For other rivers, only information about catastrophic floods is available. Table 18.2 shows the number of catastrophic floods recorded for some rivers in Spain, Italy, and France. A stationarity test (Lang et al. 1999), using different thresholds related either to a class ranking or to TABLE 18.2. The number of catastrophic floods based on historical sources recorded in various river basins in Spain, Italy, and France between 414 BC and 2000 River

Turia (1) Júcar (1) Segura (1) Guadalquivir (1) Sa Riera (1) Ter (4) Llobregat (4) Segre (4) Ebro (6) Po (2) Tiber (3) Isère (5) Drac (5)

Place

Valencia, Spain Alzira, Spain Murcia, Spain Sevilla, Spain Palma Mallorca, Spain Girona, Spain Delta, Spain Lleida, Spain Tortosa, Spain Venecia, Italy Roma, Italy Grenoble, France Grenoble, France

Period

Number of catastrophic floods

1321–1949 1328–1982 1292–1971 1301–1900 1403–1800

35 17 25 27 8

1322–2000 1315–2000 1306–2000 1355–1982 782–1990 414BC–1937 1469–1968 1373–1996

22 26 26 15 207 108 83 52

Sources: (1) Barriendos and Martin-Vide (1998); (2) Camuffo (1993); (3) Camuffo and Enzi (1996); (4) Llasat et al. (2003b); (5) Davoine et al. (2001); (6) Barriendos (1994).

a discharge value, has been applied to the above rivers. The analysis shows that extraordinary floods were not distributed uniformly over time; there were more floods in the eighteenth century, especially between 1740 and 1760. However, the catastrophic floods were homogeneously distributed in time within the period 1300– 1900, with the exception of the Llobregat floods. Longer flood chronologies selected for detailed analysis show similar behaviour patterns, with an absence of positive or negative trends and of regular cycles. However, a positive trend is seen in the extraordinary flooding of some rivers, which may be attributable to the increasing influence of human activities. The longest flood series in Table 18.2 is for the River Tiber, with 108 catastrophic floods over a period of 1,350 years (Camuffo and Enzi 1996). There are no publications about systematic series of catastrophic floods in the southern and eastern parts of the Mediterranean over a long period although good geomorphological records of flooding in recent centuries are available for Crete and north-west Greece (Chapter 11). A good deal of data exists for the Nile, but this is not a Mediterranean river. In the Medjerda River in Tunisia, ordinary floods occur every two or three years as a consequence of torrential rainfall, usually in summer and early autumn. The worst flood event recorded in Tunisia in the last 150 years was in September 1855, but other years with catastrophic or extraordinary floods have been 1902, 1905, 1906, 1907, 1937, 1969, and 1973 (Hakim 1987). The catastrophic flooding episodes recorded in the north-western Mediterranean show some oscillations of limited duration (between twenty and forty years) in which the frequency of the floods increases markedly, though with previous values regained within a few decades (Figure 18.12). This behaviour is assumed to be natural because these (climate-driven) oscillations are temporary and previous conditions are restored. The oscillations of greatest magnitude have a common presence in different basins, which goes to reinforce their general character or origin (Llasat et al. 2003b). This analysis suggests that at least three large oscillations in flood magnitude and extension took place at the end of the sixteenth century, in the last decades of the eighteenth century and in the middle decades of the nineteenth century. These results also show an attenuated reflection of other oscillations detected in central Europe, such as those of the middle years of the sixteenth century (Brázdil et al. 1999), and another coinciding with the Late Maunder Minimum (1675–1715). The series with the greatest representation of the Mediterranean littoral zones show clearly the variability in the latter years

Storms and Floods

531

1.5 30-year filter 10-year filter Flood Index

1.0

0.5

0

–0.5 1301

1401

1501

1601

1701

1801

1901

2001

Fig. 18.12. The record of catastrophic floods in Catalonia since the early fourteenth century (after Llasat et al. 2003b).

of the eighteenth century (1760–1800), an unusual period in that it coincides with a simultaneous increase in the frequency of drought episodes (Barriendos and Llasat 2003).

Wind Storms Wind storms are a key feature of the climate of the Mediterranean region. In the Adriatic area, severe windstorms have been defined as strong winds blowing over larger areas with 10-minute mean wind speeds exceeding 10 ms−1 (Tutis 2002; Ivancan-Picek and Tutis 1996). Violent Adriatic wind storms are always associated with larger mesoscale features over Europe and the Mediterranean plus the effect of the nearby mountain ranges. Numerical experiments have confirmed that the wavelength of the Alps and the Dinaric Alps has a strong influence on these windstorms. The more severe wind storms are usually produced by the bora, sirocco, and tramontana winds (Chapter 3).

The Bora The bora is a strong, gusty, north-easterly downslope wind along the eastern Adriatic coast. Recent research has shown essential differences between bora behaviour in the northern and southern Adriatic, but both are the result of multiscale (synoptic and local) factors. The most typical synoptic patterns leading to severe bora flow in the northern Adriatic are: r Anticyclonic bora type: in this case a strong anti-

cyclone north or north-east of the Alps produces strong pressure gradients over the Dinaric Alps, with flow from inland to the sea. r Frontal bora type: These are brief episodes produced by the deformation of a cold front when the cold air breaks out over the Alps—their duration and intensity is directly dependent on the amount of cold air.

r Cyclonic bora type: This involves a Mediterranean

cyclone close to the Dinaric Alps producing strong pressure gradients over the Dinaric Alps with flow from sea to inland. The Middle Adriatic or Dalmatian bora is, in most cases, the result of various physical processes including leeward mountain downslope flow and wave breaking at higher elevations. In the lower troposphere during the bora periods there is a pronounced downward NW–SE motion over Europe, which undergoes a strong acceleration upstream of the Dinaric Alps, giving rise to a lowlevel jet stream as shown in the results of the Alpine experiment, ALPEX (Kuettner 1982a, b).

The Sirocco The sirocco is a south-easterly warm and humid wind in the Adriatic Sea that originates in the Sahara. It is generally dominated by synoptic-scale events. The most severe cases are connected with a southerly upper-level airstream over the Adriatic during the presence of a Mediterranean cyclone. A low-level jet stream along the Adriatic also develops. Besides these synoptic factors, severe sirocco events are associated with a deep lee cyclone south of the Alps or a mesoscale low in the northern Adriatic (that is not detected on synoptic charts) that may draw dry and dusty air from North Africa to the Mediterranean (Chapter 14). Besides maritime traffic disruption, one of the dangerous consequences of long-lasting severe sirocco situations is an unusually high sea level that causes flooding in Venice (Camuffo 1993; Chapter 13).

The Tramontana Tramontana wind storms are produced by violent westerly or north-westerly winds over the Adriatic. They are usually produced by a cyclone’s internal circulation and the development of a low-level jet stream in the rear of a

532

María del Carmen Llasat

cold front when a cyclone is moving from Italy over the Adriatic and the Dinaric Alps and into the inland of Croatia. Wind storms are particularly hazardous for maritime traffic because of the development of high ocean-like waves on the sea surface. They are also important in the western part of the Mediterranean basin. In Spain and France they constitute the ‘mistral-cierzo’ system that can give rise to high sea waves and, on occasions, damage to trees, crops, and coastal infrastructure. Although the synoptic pattern produces strong winds from the north, mountain ranges such as the Pyrenees and the Iberian mountains in Spain, as well as the Massif Central in France, modify the main wind direction and sealevel pressure and give rise to strong local winds. The convergence between these local winds, usually over the sea, may produce a convergence line that can trigger cloudiness.

Other Severe Winds Finally, some severe storms can be associated with the ‘Levante’ wind, an easterly wind that is usually very moist. Severe Levante windstorms can arise in the Strait of Gibraltar, due to the orographic dipole that appears as a consequence of the mountain ranges of southern Spain and North Africa and the geometric features of the Alboran Sea. Other Levante events are associated with persistent rainfall that can occasionally be of high intensity and last for several days. Figure 18.13 shows the Levante wind storm and rainfall event that affected Catalonia in October 2002.

Some strong wind storms are not related to reinforcement of the usual seasonal Mediterranean winds, but instead are associated with strong cyclogenesis (Flocas and Karacostas, 1994). Generally, on such occasions, high rainfall also occurs. This was the case during the November 2001 storm, which affected Algeria and Spain. It has been classified as the most significant and violent meteorological event of the last few decades in the western Mediterranean. It arose as a consequence of the strong and rapid deepening of a surface low located off the coast of Algeria (a decrease of 8 hPa in less than 30 hours) that later moved to a position just west of the centre of the Mediterranean Sea. In Algeria, more than 200 mm of rain fell during the course of the entire event, but in the north of the Balearic Islands, in the mountains, more than 700 mm was recorded between 10 and 15 November. Wind speeds reached 35 ms−1 on these islands destroying large areas of forest (Figure 18.14) and producing sea waves more than 11 m high.

Mediterranean Tornadoes, Hailstorm Events, and Snowfalls Tornadoes Although severe storms such as tornadoes do occur in Europe, the amount of damage and loss of life that they cause is poorly documented, leading to limited awareness of the threat that they pose. For example, using

Fig. 18.13. A ‘Levante’ wind storm that affected Catalonia, 16–18 October 2002. This wind produced 4-m high waves in the port of Barcelona (photo: http://gama.am.ub. es/edrinacas).

Storms and Floods

533

Fig. 18.14. The destruction of forests in the Balearic Islands by the western Mediterranean ‘superstorm’ that hit Algeria and Spain in November 2001 (photo: courtesy of Conselleria de Medi Ambient del Govern de les Illes Balears).

the available provisional information from the European Climatology of Severe Convective Storms database (Romero et al. 2004), it has been estimated that around eighty-five significant tornadoes (of at least F2) have occurred in Europe during the period 1971–2000. The European Severe Weather Database shows that, during 2004, six tornadoes occurred in France, three in Italy, two in Malta, and one each in Tunisia, Turkey, Greece, and Slovenia. However, most European countries do not have a database of severe storm reports and many local events go unrecorded. Following the definition of Markowski (2005), ‘tornadoes are violently rotating columns of air, usually associated with a swirling cloud of debris or dust near the ground and a funnel-shaped cloud extending downward from the base of the parent cumulonimbus updraft’. They are usually associated with supercell storms, but weak tornadoes occasionally occur in non-supercellular convection. The presence of a deep, persistent mesocyclone characterizes the supercell and supercells producing ‘tornado families’ and these have a stationary evolution that can last several tens of minutes (Doswell and Burgess 1993). Although some synoptic and mesoscale conditions for tornado production are similar to those related with heavy rainfall events (such as warm air at low levels, the presence of a trough and a lowlevel jet, and so forth), the existence of a dry layer in the middle troposphere and significant wind shear are necessary. During the last fifteen years, more than 130 tornadoes have been recorded in Spain (Gayà 2004)

Fig. 18.15. A tornado recorded in Barcelona, 8 September 2005 (photo: http://gama.am.ub.es/edrinacas).

(Figure 18.15). Usually they are weak tornadoes (F0 to F3 on the Fujita scale) with wind speeds between 182 and 332 km per hour (Ayala-Carcedo and Olcina 2002). Between 1989 and 2001, eighty events with tornadoes were recorded across Spain, and fifty-five of them affected the Mediterranean region of that country (Gayà 1999) with the tornadoes in the Balearic Islands (n = 36) making the largest contribution. Twenty-three tornadoes were recorded on sixteen days between 1989 and 1997 on the Balearic Islands. Eighteen of them were of scale F0 to F1 and were concentrated between June and November (with a maximum in September) and an average path length of less than 6 km (Gayà 1998). This seasonal distribution is in accordance with general trends for Europe as a whole, where 44 per cent of tornadoes are recorded between July and September, 27 per cent from April to June, and 18 per cent from January to March. In the Spanish Mediterranean basin the largest tornadoes recorded in recent years were those of 31 August 1994, in Catalonia, and the six tornadoes recorded in the Balearic Islands on 12 September 1996. The year 2005 was also eventful, with seven tornadoes between 6 and 12 September that affected Catalonia and the Balearic Islands. The tornado recorded on 8 September 2005, which lasted only ten minutes, produced public-sector damage estimated at more than €500,000 near Barcelona (Figure 18.15),

534

María del Carmen Llasat

while at Barcelona airport 218 flights were affected by cancellations or delays, and two planes were pushed along by the wind. The previous day another tornado destroyed buildings and shops in a densely populated tourist location near Barcelona. No tornadoes with strength >F3 have been recorded in Spain in the last fifty years. By contrast, in France four violent tornadoes were recorded between 1965 and 1989 (Dessens and Snow 1989). On 15 April 2000, a tornado associated with an isolated supercell hit the Friuli region of Italy and caused extensive damage to houses, crops, trucks, and cars (Giaotti and Stel 2004). In Italy, tornadoes are not as common as heavy rainfall events, floods, landslides, and wind storms, but they are frequent and widespread enough to constitute a serious danger to people and property (Giaotti et al. 2004). In the eastern Mediterranean, Cyprus also experiences tornadoes and waterspouts, particularly in the southern regions (Sioutas et al. 2005) where the economic and social impact are significant due to the high population density. On 27 January 2003, four tornadoes affected the south-west coast of Limassol, where approximately 100 cars were hurled against one another, 5,000 buildings were damaged, about thirty people required hospital treatment, and one person was killed. Another important severe weather situation was recorded in the south of Cyprus on 22 January 2004, giving rise to violent thunderstorms, hail, and intense precipitation as well as seven tornadoes. Although preliminary analysis shows December and January as the main period for tornado and waterspout development in Cyprus, it is not uncommon for them to be recorded during the summer. The classic synoptic pattern shows a low pressure system over the south-west coast of Turkey associated with frontal activity over southern Greece and the Aegean with a strong southern flow supplying moist air at low levels with a notable low-level shear (Sioutas 2003). In Tunisia, the tornado recorded on 13 November 2004 reached F2 category, with a wind speed of 39 ms−1 . This tornado was also accompanied by heavy rain and its path length stretched for 85 km, with a NE– SW direction of movement. People were hoisted into the air by the tornado—as high as 50 m above the ground— while flooding also occurred in some places, including the city of Nabeul in the north-east.

Hurricanes ‘Medicanes’ or Mediterranean hurricanes (Emanuel 2005) are a concept that has been introduced recently into the field of severe weather in the Mediterranean

region. They are cyclonic storms that resemble tropical cyclones in satellite images and that occasionally form over the Mediterranean Sea. Emanuel (2005) shows that such storms have a small, warm-core structure and surface winds that sometimes exceed 25 ms−1 over limited areas. They develop when a deep, upper-level cut-off low moves over the region and the air under this upper low is unusually cold and humid while the underlying sea is warm.

Hail Storms Most tornadoes are accompanied by hailstorms, although the number of hailstorm events unaccompanied by a tornado is, fortunately, greater. Typically, they are produced by a super-cell system or multi-cell system (Ceperuelo et al. 2006; Sioutas and Flocas 2003) that has developed into a synoptic framework similar to the one responsible for tornadoes, but with different mesoscale features and a significant orographic contribution. In Spain, yearly hail damage is estimated at more than €100 million and during 2003 more than 50 per cent of the total damage to agriculture was due to hailstorms. Pascual (2002) identified fourteen types of weather that can produce hail in the Lleida region of north-east Spain. Analysis of the 1995–8 hailstorm database shows that forty-seven hail events were recorded in this region and 80 per cent of these events took place after 16.00 UTC with a maximum frequency in April, May, and August. In France, most hail events are usually recorded in the Atlantic region, while in Greece the main hailstorms develop in the north of the mainland. Based on an examination of the period 1976–2001, Sioutas and Flocas (2003) have classified hail days in northern Greece into seven synoptic types of atmospheric circulation. For this period, the climatological analysis shows an average of 22.3 hail days per year during the warm period of April to September. The most common synoptic types are the south-westerly flow and the shortwave trough, both being associated with intense cold advection in the upper levels and triggering of extreme instability. Piani et al. (2005) show that hailstorm hazards in central Italy are increasing and will probably continue to increase in the future. This trend has been mainly detectable in summer, since the second half of the 1970s.

Heavy Snowfall Due to the widespread orographic effects, with many Mediterranean peaks above 2,500 m in altitude, heavy snowfalls are frequent in the mountains (Chapter 12). However, the greatest impact is produced when

Storms and Floods

heavy snowfalls affect urban areas and the surrounding regions. This phenomenon occurs in winter or early spring when moist sea air impinges over the coast and littoral ranges at low levels, while cold continental air is advected at altitude. Between 14 and 15 December 2001, snow fell at sea level in Catalonia, with strong north winds (31 ms−1 ) and a maximum sea wave height of nearly 10 m. Low temperatures in the following few days froze the snow, disrupting road transport and isolating some villages. In Israel, between 16 and 18 March 1998, a very deep cyclone led to severe weather producing snowfall in Jerusalem and a sand storm. Serious damage was caused to agriculture, together with traffic delays and road closures.

The Impact of High-Magnitude Storms and Floods in the Mediterranean Region Leaving aside droughts, floods are the most dangerous meteorological hazard affecting Mediterranean countries, followed by wind storms and hail. A report presented by the European Union at Dresden on 13 October 2003 showed that the European countries most severely affected by floods in recent years were France (22 per cent of the total in Europe) and Italy (17 per cent). The same report showed that the largest number of deaths produced by floods was in Italy (38 per cent), followed by Spain (20 per cent), and France (17 per cent). The rainstorm and flooding episode of 10 June 2000 recorded in Catalonia caused material damage estimated at over €65 million and five casualties. Over the last fifty years precipitation levels of 250 mm in 24 hours have been exceeded in various storms affecting the Llobregat basin (Figure 18.1b). On 25 September 1962 that amount was recorded in less than three hours and led to 441 deaths and 374 missing persons. Between 1950 and 1999 there were more than 2,200 deaths in Spain and material damage amounting to over €301M a year (Llasat et al. 2003b). The CCS (Consorcio de Compensación de Seguros or State Insurance Clearing Company) in Spain paid out €1,574,530,945 in flood-damage compensation between 1971 and 2002, thereby accounting for almost 80 per cent of the overall amount paid out for disasters (Llasat 2004). The catastrophic floods of 22 and 23 September 1992 produced more than forty-two casualties in Vaissonla-Romaine in the Provence-Alpes-Côte d’Azur (PACA) region of southern France. During the same month, the

535

flood event that started on the 26th and ended on the 28th claimed four casualties and one missing person in France and as well as casualties in the urban area of Genoa in northern Italy, where more than US $12M was needed for urgent salvage work. Between 8 and 9 September 2002, more than 600 mm of rainfall was recorded in the Gard region (PACA) of France, although an area of 5,393 km2 recorded more than 200mm (Legrand et al. 2003). A total number of six Departments were affected with damages of €1,200M (€830M for the Department of the Gard) and twenty-three casualties were reported. Across Italy, during the twentieth century, it has been estimated that approximately 3,000 places have been affected by a flood event (Guzzetti et al. 1994). Besides floods, landslides and debris flows produced by high rainfall events are also important in the Mediterranean region, particularly in Italy (Guzzetti and Tonelli 2004; Guzzetti et al. 2005; Chapter 6). In 1998 the Sarno landslide accounted for 300 deaths in an urban area where many of the buildings had been built without official planning permission. In September 2000, a local flood event and debris flow struck a campsite in Soverato on the Calabrian coast with ten casualties. During the Valltelina (Italy) event, 407 mm of rain fell during 18–19 July producing a catastrophic flood on the Adda River and catastrophic landslides. The outcome was forty-five casualties and total losses exceeded €900,000 (Azzola and Tuia 1989). The Lombardy floods between 13 and 27 November 2002, during which there were also landslides, produced private and public sector damage to a value of €8,182,000 and €5,430,000 respectively. In the eastern and southern part of the Mediterranean basin, heavy rains are not the most frequent or dangerous meteorological hazard. In some regions, droughts, hail, and forest fires are more significant. This is the case in Croatia, where the damage produced by droughts between 1980 and 2002 has been estimated at US $2,027,922 (36.7 per cent of the total attributed to meteorological hazards), with US $1,105,604 million (20 per cent) for hailstorms, US $409,382 (7.4 per cent) for floods, and US $381,404 (6.9 per cent) for forest fires (MEDEX project). The floods recorded in Athens in October 1994 produced damage to a value of €13M for commercial and industrial properties and €1M for houses. In Israel, the floods on 21–2 February 1997 produced eleven casualties and major losses in agriculture, while from 21–7 May 1998 flooding in Turkey caused more than seventeen casualties and left 3,000 houses ruined or evacuated, with a total evaluation of losses of US $250 million. The ‘superstorm’ that affected Algeria,

536

María del Carmen Llasat

Morocco, and Spain between 9 and 13 November 2001 left more than 600 victims in its wake and thousands homeless in Algiers, with four deaths, 220,000 trees uprooted, and up to 60 per cent of sand was washed away from beaches. This storm produced over €100 million of private-sector damage in the Balearic Islands alone. The same event produced damage costing more than €62 million in Catalonia and Levante. Human impact can be significant in increasing vulnerability to flood events and in increasing the flood hazard (Ruin et al. 2008). Besides the rainfall input, which is the most important driver, other factors that can influence flood production include the nature of the catchment’s terrain and vegetation. Both the rain intercepted by the vegetation and that which infiltrates into the terrain considerably reduce the formation of hillslope runoff and, therefore, channel flows (Chapter 8). The vegetation cover can also reduce erosion, particularly in terrain with steep gradients and soft lithologies. The effects of deforestation are much greater in the Mediterranean countries than in those of central and northern Europe, given that in the latter deforested areas rapidly become covered with grass and undergrowth. In addition to this, one of the more important land use changes is the urbanization of rural lands. Flood damage has increased as a consequence of development in flood-prone areas. This is especially important at the Mediterranean coast which has seen some of the most intensive urban expansion. On the other hand, if urban development takes place in the headwaters of river basins, channel flows may increase during large flood events. Furthermore, as zones of natural storage or marshy areas are eliminated, or infiltration has diminished as a consequence of the urbanization of the soil cover, the flood hazard will increase. Although these problems affect all parts of the Mediterranean, it has been argued in this chapter that they are especially important in Catalonia (Spain), Languedoc Roussillon and PACA (France), and Liguria (Italy), as well as in some parts of Greece.

Strategies for Mitigating Meteorological Hazards The response of society to mitigating the risk of floods and severe weather in the Mediterranean region has included measures ranging from participation in worldwide international projects through to civil protection procedures put in place in each region. But the response has not been uniform in all countries. In contrast to the large number of measures implemented in Italy

and France, for example, many parts of North Africa do not even have basic flood warning protocols. In terms of research and development projects, within the wider international context the AMHY (Alpine and Mediterranean Hydrology) and MEDEX (Mediterranean Experiment) projects are noteworthy. Started in 1991, the former lies within the FRIEND (Flow Regimes from International Experiment and Network Data) project that was implemented under the auspices of UNESCO, and encompasses all the Mediterranean countries—but with particular attention to the North African countries. Its objectives include the drawing up of databases and undertaking research into heavy rainfall episodes and flooding. MEDEX started in 2000 and has the backing of the World Meteorological Organization and a chief objective of studying Mediterranean cyclogenesis with a view to improving the forecasting of severe weather and floods in the Mediterranean, as well as analysing the social impact of these risks. The project has received significant contributions from several Mediterranean countries. The IDNDR (International Decade for Natural Disaster Reduction, 1990–2000) Conference held in Grenoble in April 1999 formed the framework for the International Conference on Natural Hazards in Mountain Regions, which also included the Mediterranean, given its major mountain ranges. The conference put forward recommendations about risk awareness and developing knowledge and cooperation (Gillet and Zanolini 1999) that formed important precedents for subsequent work. This work included the necessary improvement of risk mapping by combining various methodologies and recognition of the need for a sustainable equilibrium between structural and non-structural preventive measures, as well as programmes to raise public awareness of flood hazards. Finally, each year since 1999 has seen the annual Plinius Conference on Mediterranean Storms of the European Geosciences Union. With a multidisciplinary approach, this meeting brings together scientists and technicians from various regions in the Mediterranean basin. Within the European framework, some of the projects within the Fifth Framework Programme of the European Union, such as SPHERE (Systematic, Palaeoflood and Historical data for the improvement of flood Risk Estimation), based on the analysis of palaeofloods, historical floods, and some recent floods in certain river basins in France and Spain (Tables 18.1 and 18.2) have seen important progress. Also of importance are the European Union’s INTERREG initiatives, usually involving the participation of two or three countries, as well as regional government and scientific institutions. A key

Storms and Floods

outcome of these EU projects is a comparative analysis of the flood warning systems in the three countries (Llasat et al. 2005).

Conclusions The geographical features of the Mediterranean region favour the creation of a specific Mediterranean air mass and a distinctive Mediterranean meteorology. From a natural hazards perspective, these meteorological effects can be summarized as a high concentration of cyclogenesis, heavy rainfall events and very strong local winds with sustained speeds of 20 or even 30ms−1 . As a consequence, floods and wind storms are the most common natural hazards in this region and the ones that cause the greatest economic losses. Droughts are also important in some regions, but the associated meteorological mechanisms responsible for them are completely different (Chapter 3). A climatology of surface cyclones covering the whole Mediterranean region shows that the mean frequency for the West Zone is 2,910 cyclones per year, while in the East Zone this figure is 2,248 cyclones per year. The two maximum cyclone concentrations form in the Gulf of Genoa and in the area around Cyprus. Some studies have shown the presence of a cyclonic centre in more than 70 per cent of heavy rainfall episodes in the Mediterranean area. The conceptual model outlined in this chapter explaining the development of a high rainfall event shows the presence of a long anticyclonic situation over the Mediterranean during the days preceding the storm, which favours the formation of a Mediterranean air mass. The instability is usually produced by major warm and wet advection at low levels that nourishes the air with enough water vapour to maintain the precipitation and to increase the instability. The presence of cold air in middle and high levels is not an essential factor, although when it is present it can boost thermodynamic instability. A trough or a cut-off low is usually detected in the middle and high troposphere and this advects potential vorticity and organizes the flow, favouring the entrance of air from the sea over the coastal and mountain barriers. The presence of a Mediterranean low or a convergence line organizes the differentiated air currents as well as internal low frontal boundaries. Another possible mechanism is to substitute the low internal boundary with a mountain barrier that stops and forces the ascent of the warm, moist current. The perturbation is occasionally enhanced by the entry of subtropical or Atlantic air associated with an undulation of the subtropical jet stream or Atlantic jet stream respectively. A

537

notable increase in the sea-surface temperature has been found on some occasions. If the situation remains more or less stationary, the accumulated rainfall can reach high intensities and very large amounts. This precipitation is usually associated with Mesoscale Convective Systems and Multicells, with the cluster structure being the most frequent of the MCS cases. It is not unusual for heavy rainfall caused by a single meteorological event to affect several Mediterranean countries, such as Spain, France, and Italy. Unlike the western part of the Mediterranean, where they occur in autumn, in the eastern part the main flood events commonly take place in spring. This fact is related to the different yearly distribution of the maximum rainfall as well as the different kinds of flood. This chapter has identified four kinds of flood and flood-generating mechanisms (Types 1–4) for Mediterranean catchments. When the longest possible flood series are analysed in order to evaluate their long-term evolution, a scale of event magnitude has been proposed using the flood effects on the river channel and surrounding areas. It is then possible to distinguish between ordinary stage rises and small floods, extraordinary flooding or intermediate floods, and catastrophic flooding or large floods. This last type causes severe damage or complete destruction of infrastructure. The analysis of catastrophic and extraordinary floods from documentary records shows that at least three large oscillations in magnitude and extension took place at the end of the sixteenth century, the latter decades of the eighteenth century, and the middle decades of the nineteenth century. An absence of positive or negative trends has been found for the catastrophic floods, but a positive trend is seen in the extraordinary flooding of some rivers, which may reflect the increasing influence of human activities. In addition, it can be argued that an increase in flood vulnerability, due, for example, to urbanization in flood-prone areas, is leading to more extensive damage (Table 18.1). Although wind storms are important over the entire Mediterranean, they have a major impact in the central and eastern regions. The more severe wind storms are usually produced by the bora, sirocco, and tramontana winds. Finally, some severe storms can be associated with the Levante wind, an easterly wind that is usually very moist and affects the western part of the Mediterranean.

Acknowledgements This chapter has been partially drawn up within the framework of the MEDEX WMO project. I acknowledge the support of the following for providing information: A. Jansà, A. Genovés, B.

538

María del Carmen Llasat

Ivançan-Picek, K. Lagouvardos, M. V. Sioutas, and members of the MEDEX team. My thanks to T. Estrela for the photographs in Figure 18.9 and to M. Barriendos for the information included in Table 18.2. I would especially like to thank some of my team members, M. A. Prat, T. Barrera, M. Llasat, and T. Rigo for their help in preparing the figures and obtaining information about flood damage.

References Alonso-Sarría, F., López-Bermúdez, F., and Conesa-García, C. (2002), Synoptic conditions producing extreme rainfall events along the Mediterranean Coast of the Iberian Peninsula, in L. J. Bull and M. J. Kirby (eds.), Dryland Rivers: Hydrology and Geomorphology of Semi-arid Channels. John Wiley & Sons, Chichester, 351–72. Ayala-Carcedo, F. J. and Olcina, J. (2002), Riesgos Naturales. Ariel Ciencia, Barcelona. Azzola, M. and Tuia, T. (1989), L’alluvione dell’Adda del 18–19 luglio 1987: cause, effeti e interventi, in A. Gil Olcina and A. Morales Gil (eds.), Avenidas fluviales e inundaciones en la Cuenca del Mediterráne. Institute of Geography, University of Alicante, 185–93. Barnolas, M. and Llasat, M. C. (2007), A flood geodatabase and its climatological applications: the case of Catalonia for the last century. Natural Hazards and Earth System Sciences 7: 271–81. Atencia, A., Llasat, M. C., and Rigo, T. (2008), Characterization of a Mediterranean flash flood event using rain gauges, radar, GIS and lightning data. Advances in Geosciences 17: 35–41. Barriendos, M. (1994), El clima histórico de Catalunya. Aproximación a sus características generales (ss. XV–XIX). Unpublished Ph.D. thesis, Department of Physical Geography, University of Barcelona. and Llasat, M. C. (2003), The study of climatic anomalies by means of zonality and NAO indices. The case of ‘Malda’ anomaly in the Western Mediterranean Basin (AD 1760–1800). Climatic Change 61: 191–216. and Martin-Vide, J. (1998), Secular climatic oscillations as indicated by catastrophic floods in the Spanish Mediterranean coastal area (14th–19th centuries). Climatic Change 38: 473–91. Cœur, D., Lang, M., Llasat, M. C., Naulet, R., Lemaitre, F., and Barrera, A. (2003), Stationarity analysis of historical flood series in France and Spain (14th–20th centuries). Natural Hazards and Earth System Sciences 3: 1–10. Benito, G. and Thorndycraft, V. R. (2005), Palaeoflood hydrology and its role in applied hydrological sciences, in G. Benito, T. B. M. J. Ouarda, and A. Bárdossy (eds.), Palaeofloods, hystorical data and climate variability: Applications in flood risk assessment. Journal of Hydrology 313/12: 3–15. Díez-Herrero, A., and Fernández de Villalta, M. (2003), Magnitude and frequency of flooding in the Tagus Basin (central Spain) over the last millennium. Climatic Change 58: 171–92. Brázdil, R., Glaser, R., Pfister, C., Antoine, J. M., Barriendos, M., Camuffo, D., Deutsch, M., Enzi, S., Guidoboni, E., and Rodrigo, F. S. (1999), Flood events of selected rivers of Europe in the sixteenth century. Climatic Change 43: 239–85. Brilly, M. and Polic, M. (2005), Public perception of flood risks, flood forecasting and mitigation. Natural Hazards and Earth System Science 5: 345–55.

Byers H. R. and Braham, R. R. (1949), The Thunderstorm. Report of the Thunderstorm Project. United States Department of Commerce. Washington, DC. Camuffo, D. (1993) Analysis of the sea surges at Venice from A.D. 782 to 1990. Theoretical and Applied Climatology 47: 1–14. and Enzi, S. (1996), The analysis of two bi-millennial series: Tiber and Po river floods, in P. D. Jones, R. S. Bradley, and J. Jouzel (eds.), Climatic Variations and Forcing Mechanisms of the Last 2000 Years. Springer, Berlin, 433–50. Ceperuelo, M., Llasat, M. C., García, E., López, L., and Sánchez, J. L. (2006), Study of 11 September 2004 hailstorm event using radar identification of 2D systems and 3D cells. Advances in Geosciences 7: 215–22. Chakina, N. P. and Berkovich, L. V. (1997), Diagnostic and numerical studies of a case of cyclogenesis over the Mediterranean and Black seas. INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean. Palma de Mallorca, 14–17 April 1997, Spain, 125–30. Davoine, P. A., Martin, H., Trouillon, A., Cœur, D., Lang, M., Bariendos, M., and Llasat, C. (2001), Historical flood database for the European SPHERE project: modelling of historical information, in Proceedings of the 21st Assembly of the European Geophysical Society, Nice, France. Dessens J. and Snow, J. T. (1989), Tornadoes in France. Weather and Forecasting 4: 110–32. Doswell III, C.A. and Burgess, D. W. (1993), Tornadoes and tornadic storms: a review of conceptuals models, in C. Church et al. (eds.), The Tornado: Its Structure, Dynamics, Prediction and Hazards. Geophysical Monograph 79. American Geophysical Union, Washington, DC, 161–72. Brooks, H. E., and Maddox, R. A. (1996), Flash flood forecasting: an ingredients based methodology. Weather Forecasting 11: 560–81. Emanuel, K. (2005), Genesis and maintenance of Mediterranean hurricanes. Advances in Geosciences 2: 217–20. Estrela, T., Marcuello, C., and Dimas, M. (2000), Las aguas continentales en los países mediterráneos de la Unión Europea. Report of the CEDEX. Ministerio de Fomento, Madrid. Flocas, H.A. and Karacostas, T. S. (1994), Synoptic characteristics of cyclogenesis over the Aegean Sea. International Symposium on the Life Cycle of Extratropical Cyclones. Bergen, ii. 186–91. Fossati, D. (2004), Analyse d’une situation de crise dans un cas réel. L’avant, pendant et l’après. L’exemple de Cortenova, Lecco, Italia. La comunicació del risc. Fòrum internacional de Comunicadors Rinamed, 28–30 octubre 2004, Barcelona. Frontero, P., Binder, P., Pugnaghi, S., Lombroso, L., Santangelo, R., Buzzi, A., Paccagnella, T., and Selvini, A. (1997), Versilia flood: a prefrontal supercell system embedded in weak flow impinging on near-coastal mountain range. INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, 14–17 April 1997, Palma de Mallorca, 159–67. Gallart, F., Llorens, P., Latron, J., and Regues, D. (2002), Hydrological processes and their seasonal controls in a small Mediterranean mountain catchment in the Pyrenees. Hydrology and Earth System Sciences 6: 527–37. Gayà, M. (1998), Tornados o Caps de Fibló al Mediterrani Occidental. Tethys 1. (1999), Fenómenos severos en las Islas Baleares. Parte I: frentes de racha. Parte II: tornados, in Simposio Nacional de Predicción del Instituto Nacional de Meteorología (Memorial

Storms and Floods Alfonso Ascaso). Ministerio de Medio Ambiente, INM, Madrid, 225–33. (2004), The 1886 tornado of Madrid nowadays, III European Conference on Severe Storms, 8–12 November 2004, León, Spain. Genovés, A., Jansà, A., and Estarellas, C. (1997), First Evaluation of orographic factor in Mediterranean cyclogenesis. INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, 14–17 April 1997, Palma de Mallorca, 273–82. Giaotti, D. B. and Stel, F. (2004), A multiscale observational case study of an isolated tornadic supercell in the Friuli Plain. European Conference on Severe Storms, 9–12 November 2004, León, 16. Giovannoni, M., Pucillo, A., and Stel, F. (2004), The climatology of tornadoes in Italy. European Conference on Severe Storms, 9–12 November 2004, León, 22. Gil, V. E., Genovés, A., Picornell, M. A., and Jansà, A. (2002), Automated database of cyclones from the ECMWF model: preliminary comparison between west and east Mediterranean basins. Proceedings of the 4th EGS Plinius Conference, October 2002, Mallorca. Gillet, F. and Zanolini, F. (1999), Risques naturels en montagne (Mountain natural hazards). Actes de Colloque, Grenoble, 12–14 avril 1999, France. Fresnes, Éditions CEMAGREF. Guzzetti, F. Cardinali, M., and Reichenbach, P. (1994), The AVI Project: A bibliographical and archive inventory of landslides and floods in Italy. Environmental Management 18: 623–33. Guzzetti, F. and Tonelli, G. (2004), SICI: an information system on historical landslides and floods in Italy. Natural Hazards and Earth System Sciences 4: 213–32. Stark, C. P., and Salvati, P. (2005), Evaluation of flood and landslide risk to the population of Italy. Environmental Management 36: 15–36. IPCC (2001), Climate Change 2001. Impacts, Adaptation and Vulnerability. Cambridge University Press, Cambridge. Ivancan-Picek, B. and Tutis, V. (1996), Mesoscale structures of Adriatic cyclones. Seventh Conference on Mesoscale Processes. AMS, Boston, 44–6. Jansà, A. (1997), A general view about Mediterranean meteorology: cyclones and hazardous weather. Proceedings of the INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, Palma de Mallorca. Instituto Nacional de Meteorología and Universitat de les Illes Balears, 33–42. (2002), Advances in the MEDEX implementation. Proceedings of the 4th EGS Plinius Conference, October 2002, Mallorca. Genovés, A., Campins, J., and Picornell, M. A. (1995), Mediterranean cyclones and Alpine heavy rain flood events. MAP Newsletter 3: 35–7. Genovés, A., Riosalido, R., and Carretero, O. (1996), Mesoscale cyclones vs heavy rain and MCS in the Western Mediterranean. MAP Newsletter 5: 24–5. Jansà, J. M. (1966), Meteorología del Mediterráneo Occidental. Tercer Ciclo de Conferencias, desarrollado en el Instituto Nacional de Meteorología durante el año 1964. Servicio Meteorológico Nacional, Madrid, Series A (Memorias) 43. Kuettner, J. P. (1982a), An overview of ALPEX. Ann. Meteor. 19: 3–12. (1982b) ALPEX filed phase report. GARP-ALPEX 6A, WMO. Lang, M. and Cœur, D. (2002), Flood knowledge: history, hydraulics and hydrology. Case study on three French rivers.

539

AFCRST Paris, Chinese–French Conference on Water Resources, 6–9 November 2002, Shanghai/Suzhou, 96–102. Ouarda, T., and Bobée, B. (1999), Towards operational guidelines for over-threshold modelling. Journal of Hydrology 225: 103–17. Legrand, P., Brugnot G., and Baumont, G. (2003), Rétour d’experience des inondations de septembre de 2002 dans les départaments du Gard, de l’Hérault, du Vaucluse, des Bouches de Rhône, de l’Ardèche et de la Drôme. Contribution du Groupe d’Appui et d’expertisse scientifique. (CD-ROM). Llasat, M. C. (1991), Gota fría. Col. Boixareu Universitaria 6, Barcelona. (1993), Les inondations de 1940 en Catalogne Espagnole. Les inondations semblables pendant les cinquante années suivantes, in Servei Geològic de Catalunya (ed.), L’aiguat del 40. Inundacions catastrófiques i polítiques de prevenció a la Mediterrània nord-occidental, 137–44. (2001), An objective classification of rainfall events on the basis of their convective features: application to rainfall intensity in the Northeast of Spain. International Journal of Climatology 21: 1385–400. (2004), La vulnérabilité en Catalogne. Colloque SHF. Crues Méditerranéennes, juin 2004, Nîmes. SHF, Paris, 29–34. and Puigcerver, M. (1990), Cold air pools over Europe. Meteorology and Atmospheric Physics 42: 171–7. (1992), Pluies extremes en catalogne: influence orographique et caracteristiques synoptiques. Hydrologie Continentale 7/2: 99–115. (1994), Meteorological factors associated with floods in the north-eastern part of the Iberian Peninsula. Natural Hazards 9: 81–93. and Rodriguez, R. (1992), Extreme rainfall events in Catalonia: the case of 12 November 1988. Natural. Hazards 5: 133–51. Ramis, C., and Lanza, L. (1999), Storm tracking and monitoring using objective synoptic diagnosis and cluster identification from infrared meteosat imagery: a case study. Meteorology and Atmospheric Physics 71: 139–55. Rigo, T., and Montes, J. M. (2000), Orographic role in the temporal and spatial distribution of precipitation. The case of the internal basins of Catalonia (Spain). Proceedings of the EGS Plinius Conference, 14–16 October, Maratea, Italy, 41–55. and Barriendos, M. (2003a), The ‘Montserrat-2000’ flash-flood event: a comparison with the floods in the Northeastern Iberian Peninsula since the 14th century. International Journal of Climatology 23: 453–69. Barriendos, M., Barrera, A., and Rigo, T. (2003b), Climatological analysis of flood frequency in the Ter, Segre and Llobregat basins from the 14th to 20th century, in Thorndycraft et al. (2003) Palaeofloods, Historical Data and Climatic Variability. Applications in Flood Risk Assessment. CSIC, Madrid. 275–80. Barrera, A., and Rigo, T. (2005), Floods in Catalonia (NE Spain) since the 14th century, in G. Benito, T. B. M. J. Ouarda, and A. Bárdossy (eds.), Palaeofloods, hystorical data and climate variability: Applications in flood risk assessment. Journal of Hydrology 313/1–2: 32–47. Martín, F., and Barrera, A. (2007), From the concept of ‘kaltluftropfen’ (cold air pool) to the cut-off low. The case of September 1971 in Spain as an example of their role in heavy rainfalls. Meteorology and Atmospheric Physics 96: 43–60. Llasat Botija, M., Llasat, M. C., Créton-Cazanave, C., Lutoff, C., Ducrocq, V. (2005), Comparaison des systèmes d’alerte

540

María del Carmen Llasat

opérationnels.(Action 1.10: Analyse et comparaison entre les différents systèmes d’alerte actuellement opérationnels auprès des partenaires. Définition des exigences et des liens pour la réalisation de systèmes d’alerte opérationnels), Projet AMPHORE. FBG, University of Barcelona. López-Bermúdez, F., Conesa-García, C., and Alonso-Sarría, F. (2002), Floods: magnitude and frequency in ephemeral streams of the Spanish Mediterranean region, in L. J. Bull and M. J. Kirby (eds.), Dryland Rivers: Hydrology and Geomosphology of Semi-arid Channels. John Wiley & Sons, Chichester, 329–50. Maas, G. S. and Macklin, M. G. (2002), The impact of recent climate change on flooding and sediment supply within a Mediterranean mountain catchment, southwestern Crete, Greece. Earth Surface Processes and Landforms 27: 1087–105. Markowski, P. (2005), Atmospheric Convection Course. 18–22 July 2005, Udine, Italy. Martín León, F. (2003), Las gotas frías/DANAs. Ideas y conceptos básicos. Nota técnica STAP 38. Instituto Nacional de Meteorología. Ministerio de Medio Ambiente, Spain. Milelli, M., Llasat, M. C., and Ducroq, V. (2006), The cases of June 2000, November 2002 and September 2002 as examples of Mediterranean floods. Natural Hazards and Earth System Sciences 6: 271–84. Palmieri, S. and Clericci, R. (1992), Caratteristiche Meteorologiche degli Eventi Estremi. Gruppo Nazionale per la difesa dalle catastrofi idrogeogiche. Report 89. CNR. Presidenza del Consiglio dei Ministri. Dipartimento della Protezione Civile. Genova, Italy. Pascual, R. (2002), Estudio de las granizadas en el llano de Lleida, Nota Técnica 3, Instituto Nacional de Meteorologia. CMT Catalonia, Barcelona. Pavese, M. P., Banzon, V., Colacino, M., Gregori, G. P., and Pasqua, M. (1992), Three historical data series on floods and anomalous climatic events in Italy, in R. S. Bradley and P. D. Jones (eds.), Climate Since A.D. 1500. Routledge, London, 155–70. Piani, F., Crisci, A., de Chiara, G., Maracchi, G., and Meneguzzo, F. (2005), Recent trends and climatic perspectives of hailstorm frequency and intensity in Tuscany and Central Italy. Natural Hazards and Earth System Sciences 5: 217–24. Radinovic, Y. (1997), The basic concept on the methodologies of Mediterranean cyclones and adverse weather phenomena studies. Proceedings of the INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, 14–17 April 1997, Palma de Mallorca, 45. Ramis, C., Llasat, M. C., Genovés, A., and Jansà, A. (1994), The October 1987 floods in Catalonia: synoptic and mesoscale mechanisms. Meteorological Applications 1: 337–50. Alonso, S., and Llasat, M. C. (1995), A comparative study between two cases of extreme rainfall events in Catalonia. Surveys in Geophysics 16: 141–61. Rigo, T. and Llasat, M. C. (2003), Features of convective systems in the NW of the Mediterranean Sea. Proceedings of 5th EGU

Plinius Conference on Mediterranean Storms 1–3 October 2003, Ajaccio, France. European Geosciences Union. (2004), A methodology for the classification of convective structures using meteorological radar: application to heavy rainfall events on the Mediterranean coast of the Iberian Peninsula. Natural Hazards and Earth System Sciences 4/1: 59–68 Riosalido, R., Ferraz, J., Alvarez, E., Cansado, A., Martín, F., Elízaga, F., Camacho, J. L., and Martín, A. (1997), A flash flood event in the Spanish Pyrenees: the Biescas case. INM/WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, April 1997, Palma de Mallorca, 151–8. Romero, R., Sumner, G., Ramis, C., and Genovés, A. (1999), A classification of the atmospheric circulation patterns producing significant daily rainfall in the Spanish Mediterranean Area. International Journal of Climatology 19: 765–85. Gayà, M., and Doswell, Ch. A. (2004), European climatology of severe convective storm environmental parameters: a test for significant tornado events. European Conference on Severe Storms, 9–12 November 2004, León, Spain, 9. Ruin, I., Creutin, J.-D., Anquetin, S., and Lutoff, C. (2008), Human exposure to flash floods: relation between flood parameters and human vulnerability during a storm of September 2002 in Southern France. Journal of Hydrology 361: 199–213. Scherhag, R. (1948), Wetteranalyse und Wetterprognose, in Neue Methoden des Wetteranalyse un Wetterprognose. Springer, Berlin, 227–35. Sioutas, M. V. (2003), Tornadoes and waterspouts in Greece. Atmospheric Research 67/8: 645–56. and Flocas, H. A. (2003), Hailstorms in Northern Greece: synoptic patterns and thermodynamic environment. Theoretical and Applied Climatology 75: 189–202. Doe, R., Michelides, S., Christodoulu, M., and Robins, R. (2005), Meteorological conditions contributing to the development of severe tornadoes in southern Cyprus. Weather, 60. Steiner, M., Houze Jr., R. A., and Yuter, S. (1995), Climatological characterization of three-dimensional storm structure from operational radar and rain gauge data. Journal of Applied. Meteorology 34/9: 1978–2007. Taramasso, A., Gabellani, S., and Parodi, A. (2005), An operational flash-flood forecasting chain applied to the test cases of the EU project HYDROPTIMET. Natural Hazards and Earth System Sciences 5: 703–10. Thorndycraft, V. R., Benito, G., Barriendos, M., and Llasat, M. C. (eds.) (2003), Palaeofloods, Historical Data and Climatic Variability. Applications in Flood Risk Assessment. CSIC, Madrid. Tutis, V. (2002), Violent Adriatic windstorms. Mediterranean storms. Proceedings of the 4th EGS Plinius Conference, October 2002, Mallorca. WMO (1995), The Twelfth World Meteorological Congress, the General Summary, para. 3.3.2.4.

This chapter should be cited as follows Llasat, M. C. (2009), High magnitude storms and floods, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 513–540.

19

Wildfires Francisco Lloret, Josep Piñol, and Marc Castellnou

Introduction Fire is currently recognized as one of the major natural hazards of the Mediterranean basin. In an average year the total burnt area in the whole basin is around 600,000 hectares, the product of approximately 50,000 fires. The estimated annual cost is around 775 million Euros (FAO 2001). Official data on casualties due to fires are often not available, but, for example, seventy-nine people have been killed directly by fire in Portugal since 1966 and fifty in Catalonia (northeast Spain) since 1970. Fire is commonly considered to be a key component of the dynamics of Mediterranean ecosystems (Chapters 7 and 23). Long dry periods, usually in summer, and vegetation assemblages that produce large amounts of standing branches and debris, are the main factors promoting the propagation of fires. These characteristics are common to other regions of the world with a similar climate and vegetation structure including California, central Chile, South Africa, and south-western Australia. Fire is a common occurrence and a significant natural hazard in all these regions. Although initially a natural phenomenon, during the course of the Holocene human activity has become an increasingly powerful driver of fires (Chapter 9). Prevention of wildfires is now one of the top priorities of the forestry and environmental agencies across the Mediterranean region because of the huge extent of the burned surface area, the high expenditure on both fire prevention and fire fighting, and the impacts in terms of both human life and property. The development of models to investigate the relative roles of extreme weather conditions and fire suppression policies in the generation of large fires is a key area of research (Piñol et al. 2007).

Number of Fires and Extent of the Burnt Area The pattern of fire occurrence is not uniform across the Mediterranean basin and orders of magnitude differences appear (Figure 19.1a). When standardized to the forested area of each country, the average burnt area exceeds 103 ha per 103 km2 per year in Greece, Israel, Italy, Algeria, Portugal, and Spain. This means that, on average, more than 1 per cent of the forested area is burnt in these countries each year (Figure 19.1b). Most of the fires occur in summer, although the proportion of fires occurring in other seasons may also be significant. In contrast, less than 100 ha per 103 km2 per year are burnt in Albania, Bosnia-Herzegovina, Slovenia, and Morocco. When making such comparisons, however, it is important to consider the origin and reliability of the fire data. First, information sources mostly come from official agencies of the respective countries and they may be processed following different criteria. The definition of a forest fire, for example, may vary from country to country. The length of the records is also variable and in several countries data are only available for the period since the 1990s. Second, local historical and political factors (such as the recent conflicts in the former Yugoslavia) can influence both the official record and the dynamics of fire occurrence in the landscape. Third, the stochastic nature of fire means that exceptional events can result in high annual values. This seems to be the case in Algeria, where the burnt surface in 1983 was more than 220,000 ha—which is equivalent to 10 per cent of the total forested area of the country. The structure of fire size classes will be discussed below. All these factors make it difficult to establish a clear spatial pattern across the Mediterranean region. While

542

Francisco Lloret, Josep Piñol, and Marc Castellnou Number of fires/year/1,000 km2

(a)

1,000–5,000 500–1,000 100–500 10–100 0–10

Hectares burnt/year/1,000 km2

(b)

2,000–5,000 1,000–2,000 500–1,000 50–500 0–50

0

500 km

Fig. 19.1. Fires in Mediterranean basin countries weighted by the respective forested land shown as (a) the number of fires per year per 103 km2 , and (b) the burnt area in hectares per year per 103 km2 . The period considered ranges from four years in the case of Lebanon (1996–9) to thirty-nine years (1961–99) in the case of Morocco (after FAO 2001).

Wildfires

Number of fires per year

105

(a)

EU5

Spain

104

Morocco

103

102 106

EU5

(b) Spain

Area burnt (hectares) per year

the burnt surface area is high in some of the more economically developed countries, it may also be dramatic in some less-developed ones. Low values appear in North Africa, in the eastern Mediterranean, and in the former Yugoslavia (apart from Croatia, which has a long coastline with a typical Mediterranean climate). The low values for France are the result of the high proportion of non-Mediterranean regions within the country. When only the fifteen Mediterranean departments of France are considered, the burnt area increases from 158 to 861 103 ha per 103 km2 of forested area per year (data from the French Administration and see the Prométhée database at <www.promethee.com>). It is often assumed that a steady rise in the fire problem has been apparent in the Mediterranean in recent years, but the pattern is somewhat more complex. First, a distinction should be made between the number of fires and area of burnt surface. The number of fires is of interest since it provides insights into the trend for the number of ignitions, and because important resources are allocated to the detection and early suppression of fires. If we consider the evolution of the number of fires in a country, such as Spain, with records since 1961, we observe a significant increase in the number of fires and this is well documented in the scientific literature (Piñol et al. 1998; Moreno et al. 1998). However, the number of fires seems to have stabilized over the last ten years (Figure 19.2a). The same pattern is observed in other countries, such as Morocco or in the bulk of the European Mediterranean (Greece, Italy, France, Spain, and Portugal) (Figure 19.2a). This stabilization may be the result of the policies of fire prevention and early extinction strategies that are prevalent in these countries. Burnt surface area is a better estimator of the fire phenomenon itself and of its economic and ecosystem impact. In this case, a plateau seems to have been reached, and no trend is apparent (Figure 19.2b). Variability between years for a given region is, however, high and normally not synchronous across the whole basin. For example, in Portugal in 2003, around 420,000 ha were burnt, equivalent to 11.5 per cent of the forested area. In the same year, in Italy around 90,000 ha were burnt (0.9 per cent of the forested area), but in Greece only 3,400 ha (0.09 per cent of the forested area). The stochastic nature of fires and the marked interannual fluctuations in climate—which is typical of Mediterranean regions (Chapter 3)—both contribute to create this pattern of variability. In some countries, such as Morocco, the burnt surface area is systematically lower (Figure 19.2b), even after applying a weighting for the total forested area (Figure 19.1b). Since the climatic conditions are even more conducive to fire in this

543

105 Morocco 104

103

102 1960

1970

1980

1990

2000

Year

Fig. 19.2. Temporal variation in (a) the number of fires per year, and (b) the area burnt in Spain, Morocco, and in five of the Mediterranean countries of the EU (Greece, Italy, France, Spain, and Portugal). When one considers the entire period, there is a significant linear increase in fire frequency in Spain of 559 fires per year (r = 0.92; p < 0.001), in Morocco of 5.3 fires per year (r = 0.60; p < 0.001), and in the Mediterranean EU of 1892 fires per year (r = 0.80; p < 0.001). However, since about 1990 the number of fires in Spain and in the Mediterranean EU seems to have stabilized. The area burnt increased significantly in Spain between 1961 and 1989 (10,465 ha per year; r = 0.69; p < 0.001) and then stabilized. No temporal trend was significant for the area burnt in Morocco and in the Mediterranean EU.

part of the Mediterranean, such a pattern may be the result of depleted fuel loads related to socio-economic factors.

Fire Regime: Fire Size and Return Times A key factor contributing to the marked spatial and temporal variability of fires is the extreme skewness of the distribution of fire magnitude. Most of the burnt area

544

Francisco Lloret, Josep Piñol, and Marc Castellnou 60 Number of fires Area burnt

Percentage of fires Percentage of area burnt

50

40

30

20

10

0

<1

1–10

10–100 100–1,000 Size of fire (ha)

>1,000

Fig. 19.3. The proportion (per cent) of the number of fires and of the area burnt in relation to fire size in the fifteen Mediterranean departments of France for the period 1973–2004. Source: <www.promethee.com>.

is accounted for by a very small proportion of fires. These major fires can achieve catastrophic dimensions in terms of their extent and intensity as witnessed, for example, in the Peloponnese region of Greece in the summer of 2007. This can be clearly seen, for example, using the fire record of Mediterranean France (Figure 19.3). The largest 0.15 per cent of fires (>1,000 ha) burnt 39 per cent of the total area, the largest 1.4 per cent of fires (>100 ha) burnt 73 per cent of the total area, and the largest 7.4 per cent of fires (>10 ha) burnt 90 per cent of the total area. This is an almost universal pattern for any region with wildfires—including non-Mediterranean environments—and it has been well documented for the whole of the Mediterranean basin (Ricotta et al. 2001). This pattern is of great importance for managers in order to establish strategies to avoid these extreme events, when uncontrolled wildfires cause the most impact on properties and natural resources and threaten human lives. In Portugal, two periods of extreme fire hazard in the first half of August and the second week of September 2003 resulted in simultaneous fires that are estimated to have burnt more than 440,000 ha and killed twenty-one people (European Commission 2004; Viegas 2004) (Figure 19.4). In July 1994, again during extreme conditions of high temperature and winds, seventeen fires of more than 1,000 ha—many of them occurring simultaneously—burned 170,000 ha

and produced twenty-two casualties in Spain (Moreno et al. 1998). The occurrence of simultaneous fires is an additional cause of increased fire size, as fire-extinction resources become overstretched. These extreme episodes also introduce a source of uncertainty in the temporal evolution of the burnt surface, as they can destroy any stabilization tendency. In Portugal, the annual burnt area fluctuated steadily around 50,000 to 150,000 ha prior to the extreme event of 2003 (European Commission 2004). As stated above, burnt area estimates are spatially variable because of the different methodologies used in each country (Vélez 2000). In many cases the analysis of fire origins and consequences relies on accurate mapping of the extent of the burnt area. Many research initiatives have shown the value of satellite-based remote sensing to obtain reliable and objective estimates of these areas across large territories (Salvador et al. 2000; Chuvieco et al. 2002). The spectral properties of the burnt areas determine the reliability of the available methodologies (Pereira et al. 1999). The visible spectral domain (e.g. from Landsat Thematic Mapper TM and National Oceanic and Atmospheric Administration’s Advanced Very High Resolution Radiometer NOAA/AVHRR) has been extensively used for mapping burnt areas. In this spectral domain, recently burnt areas are relatively dark, but they can be difficult to discriminate from other dark land covers such as water bodies and some soil types. However, burnt areas can be distinguished because they tend to produce post-fire signals resulting from vegetation recovery. In the near-infrared (NIR) spectral region, the signal of recent fires is strongest—particularly when large amounts of charcoal are produced—and it has been widely used for fire mapping. A common limitation of these methodologies is that only fires above a given threshold (30–100 ha) can be accurately detected and mapped. However, since most of the burnt area is the result of large fires, the unaccounted for surface is a small proportion of the total and can be estimated by regression (European Commission 2004). Since 2001 the European Commission routinely provides annual estimates of the area burnt in fires larger than 50 ha in Portugal, Spain, France, Italy, and Greece based on WiFS or MODIS imagery (European Forest Fire Damage Assessment System, EFFDAS; European Commission 2004; Figure 19.4). This information may be used to establish several parameters of the fire regime related to the temporal sequence of fires. The interval between fires is the elapsed time between two consecutive fires in a geographical location. Fire period is the length of time needed for a given area to experience a fire; it is the inverse of fire

Wildfires

545

Fig. 19.4. The area burnt by forest fires in Portugal during 2003. Burnt areas are shown in black over a grey-composite of satellite images (European Commission 2004). The 2003 fire season in Portugal was the worst ever recorded, with over 400,000 hectares burnt, a figure more than double that of the previous highest year (1991). There were twenty-one casualties in 2003 (Viegas 2004).

frequency, that is, the number of fires occurring in a given area per unit of time. Unfortunately, the length of the survey period is generally too short to establish such parameters accurately, particularly when temporal and spatial variability is high. Some researchers have found that fires are not randomly distributed, but

that vegetation, climate, topography, and human activity determine the spatial pattern of fires (Vázquez and Moreno 2001; Díaz-Delgado et al. 2004a ). These findings agree with rough estimates obtained from administrative sources and offer the advantage of covering larger regions (Moreno et al. 1998).

546

Francisco Lloret, Josep Piñol, and Marc Castellnou

Díaz-Delgado et al. (2004b) used Landsat MSS images to produce yearly maps of fire occurrence using the NDVI (Normalized Difference Vegetation Index) to describe the fire regime in Catalonia (Spain) from 1975 to 1998. They found that the estimated mean interval of fire occurrence in landscape units of c.8 km2 was 5.5 years with a fire period of 133 years, a value similar to the 100 years estimated for the whole of Spain from administrative records (Moreno et al. 1998). Fire hazard appears to increase asymptotically before reaching a plateau at around 15–16 years after a fire. However, it is important to stress that recurrent fires occur quite often even in recently burnt areas, since 14 per cent of the burnt area experienced at least one additional fire during the 24-year period, with a maximum of up to five fires in areas with strong winds.

Historical Context The Holocene Since the physical determinants of fire (fuel, heat, and oxygen) are basically independent of human activity, provided natural sources of ignition occur, wildfires should have existed before human settlement. This seems to be the case for regions such as California, Australia, and South Africa where lightning storms are not uncommon and wildfires have occurred since prehistoric times (Kemp 1981; Keeley 1992; Keeley et al. 1999). Lightning storms also occur in the Mediterranean basin (Vázquez and Moreno 1998), although the surface area burnt by this kind of fire is now commonly much reduced compared with earlier periods of the Holocene. Lightning fires usually start under moist conditions—they therefore behave as low-intensity fires that, in the past, could have continued until the onset of heavy rain. Records of charcoal particles from lake sediment sequences (often conducted in association with pollenbased reconstructions of vegetation change) constitute an important indicator of past regional fire occurrence (MacDonald et al. 1991). Parts of the Mediterranean have well-preserved, high-resolution lake sediment records (Chapter 9) and some good records of Holocene fire history are beginning to emerge (e.g. Sadori and Giardini 2007). In these studies, the charcoal counts often provide data on large, high-intensity fires, with much less information on low-intensity fires. Most of the available information comes from the western part of the Mediterranean basin, showing that fires were occasionally present in the Iberian Peninsula during the last cold stage (c.30,000 to 15,000 BP) (Carrión 2002).

However, a significant increase in charcoal deposition does not arise until the establishment of much warmer and seasonally dry conditions, typical of the Mediterranean climate (Jalut et al. 2000). The precise dates may be different between localities, but from c.6,000 BP there is a clear increase in charcoal in such records. This pattern is consistent in several localities in the southern (from c.4,200 BP in Sierra de Gador) (Carrión et al. 2003), south-eastern (from c.5,000 BP in Sierra de Segura) (Carrión 2002), and the central Iberian Peninsula (from c.5,000 BP) (Franco-Múgica et al. 2005). In fact, the use of fire by people as a driver of landscape transformation is documented in the mid-Holocene in French sub-mediterranean mountains, where Pinus sylvestris forests moved to the present-day rangeland during the period 4,800–3,000 BP (Quilès et al. 2002). In all these records, fire occurrence increases up to the present day, believed to be largely as a result of human activities. This trend is well illustrated by the charcoal peaks obtained in the eastern Iberian Peninsula corresponding to periods of intense landscape transformation during the Roman agricultural expansion (c.2,000 BP), the medieval increase in grazing area (c.1,500–1,300 and 1,000 BP), and, more recently, after the eighteenth-century population rise (Riera-Mora and Esteban-Amat 1994). Overall, although evidence for fires is present, it seems that biomass burning was low in the western Mediterranean basin during the early Holocene (Carcaillet et al. 2002) and that human activity has been a major driver of fires since about 2,000 BP. Data from other parts of the world with Mediterranean-type climates suggest that fire has long been a major factor in shaping these Mediterranean ecosystems. In south-western Australia, for example, high-resolution records from the Pliocene have shown the existence of fires at intervals of 5–13 years prior to the establishment of human populations (Atahan et al. 2004). In the same region, there is evidence for fire throughout the Holocene, but with regional and temporal variability (Dodson 2001). In southeastern Australia, charcoal has been reported to be strongly associated with sclerophyllous vegetation during the last interglacial. Once established, the fire regime demonstrates a reasonably constant pattern of occurrence even during the last glacial period and the Holocene, suggesting that humans may have contributed to this regular pattern of fire occurrence (Singh et al. 1981). In California, by contrast, charcoal deposition indicates that the regime of large fires has not changed dramatically over the last 500 years, showing a strong relationship with climate. Here they can be assumed to be part of the natural fire regime of

Wildfires

the region, where smaller fires would also be present (Mensing et al. 1999). Overall, it appears that in recent millennia, in these other Mediterranean regions, fire regime has not experienced such strong fluctuations over the fine temporal scales that are evident in the Mediterranean basin, where historical events have produced a changing pattern of fire occurrence. Fire recurrence intervals in these regions are estimated at 4–45 years in the South African fynboss (van Wilgen et al. 1992), 30–60 years in the Californian chaparral (Keeley 1992), and 3–13 years in the Australian heath lands (Atahan et al. 2004).

Historical Times and Data Sources For the historical period features such as fire scars on trees can provide information on low- or mediumintensity fires (Figure 19.5). Although this information is often very scarce, in the north-east Iberian Peninsula low-intensity fires have been documented at a frequency of around twenty-five years since the beginning of the eighteenth century (Pellisa, personal communication).

Fig. 19.5. A fire scar at the base of a pine tree trunk in Catalonia, north-east Spain. This feature is indicative of a low-intensity fire that did not burn the forest canopy (photo: Marc Santandreu).

547

This frequency has permitted the persistence of a mature structure of Pinus nigra forests by allowing tree survival but at the same time thinning the dense regeneration (Castellnou et al. 2002a ). Written sources of information are also available, including occasional news of fires, or even detailed records. This is the case in the archives of the city of Tortosa in Catalonia for the fourteenth and fifteenth centuries. The expenses of the city were recorded in detailed accounts, including the payments to people who fought fires in the surrounding forests. It is possible, therefore, to reconstruct the frequency of fires during this period. Although the comparison between historical periods has some significant shortcomings due to differences in the accuracy of the recorded information, the results suggest the occurrence of a number of fires of the same order of magnitude as in the last decades of the twentieth century (Lloret and Marí 2001). The size distribution of fires— with many small fires and very large ones—is also comparable between these periods. Occasional information from travellers provides only fragmentary data that do not allow the importance of fires to be established. These observations often show a scenario where burning was quite common for improving pastures, while in other cases the records describe extraordinary events. Legislation that banned such fires in forested areas is often found in several areas of the Mediterranean (including Sardinia, Catalonia, Castile, and Crete) (Grove and Rackham 2001) that were associated with the exploitation of forest goods. In summary, it appears that two different attitudes to fire coexisted in many parts of the Mediterranean until the twentieth century. One approach involved the use of deliberate fire ignition to obtain forest clearings and managed pastures, while the other was based on fire prevention by banning burning activities and prosecuting those who carried them out. In some cases it resulted in active fire-fighting. Landownership and historical circumstances would control the balance of these activities. These, combined with climate conditions, would determine the historical pattern of fire occurrence in a given area. Although a detailed reconstruction of fire history is not available for any country, it seems that there was not a regular pattern of fires across time. Instead, periods of variable burning incidence took place prior to a steep increase in the last two centuries (Riera-Mora and Esteban-Amat 1994). In the late nineteenth and early twentieth centuries, administrative and journal sources provide more regular sources of written information. From Spanish administrative records, we know that at the end of the nineteenth century, fire was frequent in Spanish forests,

548

Francisco Lloret, Josep Piñol, and Marc Castellnou

with a similar frequency and size structure (most of the surface is burned by a few fires) to the present (Araque 1999). Some localities were particularly prone to burning, probably due to fire ignited by traditional rural communities. In fact, in this period, ideas on fire prevention and fire-fighting were introduced by forest engineers in a rural society where fire was commonly used. While catastrophic fires occasionally occurred at the beginning of the twentieth century, this conflict remained, and eventually fire-fighting strategies became dominant.

Determinants of Fire Regime: Climate, Vegetation, and Social Context Combustion is an exothermic reaction determined by three main factors: fuel, heat, and oxygen. Although the chemical processes involved in fire are relatively simple, the controls on fire regime are rather more complex, since historical and biological processes are also involved, as shown in Figure 19.6. Fuel provided by Mediterranean vegetation is particularly prone to produce high-intensity and fast-spreading fires. Mediterranean plants usually accumulate dead branches and leaves in standing shoots and on the ground. Low decomposition rates associated with water scarcity during long periods enhance fuel accumulation. In addition, this material shows a high surface area to volume ratio, which is a consequence of the reduction of the evaporative surface area on the leaves of Mediterranean plants, and of the common occurrence of defensive structures in the bark, shoots, and leaves (Chapter 7). Such a high ratio has two consequences. First, water content largely fluctuates according to drought periods. As the water content decreases in fuel, less heat is needed to propagate the combustion reaction. Second, the chemical reaction is facilitated because more oxygen per surface unit is in contact with fuels. The chemical characteristics of Mediterranean vegetation also contribute to promote the spread of fire, as volatile compounds are abundant. These compounds, mainly terpenes, are extremely flammable and have relatively low boiling points, thus further promoting the spread of the burning front. The Mediterranean climate is characterized by strong seasonality, with high temperatures and low air humidity in summer (Chapter 3). At this time of year the water content of fuel is very low, which facilitates the ignition and propagation of fire. Also, local winds increase both fuel desiccation and fire propagation, creating

favourable conditions for large and intense fire events. Accordingly, climatic records from eastern Spain show that in wet summers the area burned is lower than in dry summers (Pausas 2004). However, climatic conditions do not remain constant throughout the whole summer. In fact, a wildfire is a stochastic event that is usually concentrated within a few days. This may produce difficulties for attempts to correlate seasonal or annual average climatic parameters with fire occurrence in terms of the number of ignitions and amount of burnt area. Accordingly, the yearly number of days with extreme climatic conditions correlates well with the annual number and area of fires (Piñol et al. 1998). Climatic determinants of fire may, however, be rather more complex. The growth of vegetation, such as herbs and grasses, enhanced by abundant rainfall, may result in more fuel being available to burn in future summers. Indeed, Pausas (2004) has found a positive correlation between summer rainfall and summer burned area for a time lag of two years in eastern Spain. A particular challenge is to assess the significance of very recent climatic changes on fire regime. Interestingly, the recent rise in temperatures (about 0.35 and 0.27◦ C for annual and summer temperatures per decade respectively in eastern Spain) does not correlate with fire occurrence (Pausas 2004). However, the combined effect of rainfall and temperature produces a different pattern. Hence, in the same region the number of summer days with high climatic fire risk (that is, combining high temperatures and low humidity) has increased significantly during the twentieth century correlating with the rise in both the number and extent of fires (Piñol et al. 1998; Figure 19.7). The combination of natural factors promoting fire occurrence has been enhanced by human activity. Apart from the effect of human activity on climate change, humans directly modify the fire regime by creating ignitions or by fighting fires. The natural source of ignition is lightning. Although geographically variable, it accounts for about 10 per cent of wildfires (Vélez 2000). Lightning-caused fires contrast with human-caused fires, since the latter are clustered in certain geographical areas, concentrated in summer, and show smaller maximum sizes (Vázquez and Moreno 1998). Humans introduced fire as a land management tool in the Mediterranean region thousands of years ago when fire was deliberately used to improve grazing or to facilitate clearing. Nowadays most such practices are declining in the northern part of the basin, but rural traditions continue in large parts of Mediterranean North Africa. However, agricultural practices remain a major source of ignitions, even in

Wildfires

(a)

549

(b) Adverse meteorology

Ignitions +

+

+

+ Area burnt

Area burnt

– –

+

Adverse meteorology

Ignitions

Fuel build-up

–

+ Fire suppression

Fuel build-up

+

Fire suppression

Fig. 19.6. A conceptual model of factors influencing the area burnt in a region showing (a) without interaction between the factors, and (b) with the main interactions that influence fuel build-up.

Number of days

90

(a)

60

30

0 1940 106

1950

1960

1970 Year

1980

1990

(b)

Area burnt (ha)

105

104

103

102 0

30 60 Number of days

90

Fig. 19.7. (a) Changes in summer climatic fire risk in Catalonia (north-east Spain) between 1940 and 2000 (r = 0.57, p < 0.01), and (b) the burnt area in relation to the number of days with high climatic fire risk (exponential fit, r = 0.88, p < 0.01). Fire risk is estimated as the number of days exceeding a predetermined value of the climatic I87mod index. This index considers daily maximum temperature, minimum relative humidity, and maximum wind speed (see Piñol et al. 1998 for calculation details). (Redrawn from Piñol et al. 1998).

the European Mediterranean (Vélez 2000), while new sources have appeared, including those related to economic development (e.g. traffic, electrical power lines, rubbish dumping) and recreational uses. For instance, during the period 1968–98, most wildfires in eastern Spain started on Sundays (Terradas et al. 1998). Indeed a positive correlation between population density and number of fires has been observed (Terradas et al. 1998). A major effect of human activity in fire regime modification is related to socio-economic factors leading to profound changes in land use patterns (Moreira et al. 2001; Tàbara et al. 2003). The abandonment of traditional farming in many countries of the northern Mediterranean basin has taken place during the course of the twentieth century. Consequently, fuel load has grown as a result of vegetation succession. Paradoxically, this trend is enhanced by those fire-fighting policies that prevent the fuel reduction effect produced by small fires. The increase of wildfires in the last decades of the twentieth century is the logical result of the combination of huge fuel loads, increasing climatic risk, and multiple ignition sources. It is often advocated that fuel accumulation may be minimized by grazing or clearing. These were common practices a few decades ago, but they are no longer economically viable in most areas. Public funds to support these activities over vast areas are not feasible. By contrast, in North Africa, local populations keep fuel levels low by grazing or by collecting it for heating—as was common practice decades ago in the European Mediterranean countries. While wildfires seem to threaten the fate of forests in the northern Mediterranean, overexploitation tends to limit forest recovery in the south.

550

Francisco Lloret, Josep Piñol, and Marc Castellnou

Wildfire Hazard and the Impacts on Ecosystems Fire Hazard Fire hazard or risk has a structural component and a time-varying component. The structural risk arises from topography, vegetation characteristics, and some human-related factors (Chuvieco et al. 1999a). The time-varying risk comprises the relatively slow change in the water content of fuels and a much faster change in meteorological conditions such as wind direction and speed and air temperature and humidity. Structural risk is usually known from topographic and vegetation maps. Of these, the most useful are those that classify the vegetation within the so-called fuel models (Rothermel 1983). These classify vegetation types according to their most important characteristics that affect the spread of fire. Fuel model maps are usually obtained from a combination of field survey and photointerpretation, but the process can be speeded up by using satellite-based imagery (Riaño et al. 2002). Unless an important fire occurs, these maps only need to be updated every few years. The fuel moisture content (FMC) is estimated differently for dead and live fuels. In general, dead FMC is estimated from meteorological variables (Viney 1991), as it is the result of a simple physical process, that is, the exchange of water between air and the dead leaf or twig. As the estimation of the FMC of living fuels is much more difficult, this variable is often measured on selected locations and species (Viegas et al. 1992; Viegas et al. 2001). When it is desirable to have a map of living FMC it is necessary to extrapolate the key measurements using some kind of model based on variables routinely measured in meteorological stations (Viegas et al. 2001). Alternatively, living plus dead FMC can be estimated by satellite remote sensing, although the estimate is usually better for grasslands than for shrublands (Chuvieco et al. 1999). The most sensitive variables for FMC estimation are based on short-wave infrared bands, and the combination of vegetation indices and surface temperature (Chuvieco et al. 1999b). At an operational level, national fire agencies rely on the calculation of meteorological indices of fire danger. In the past each country used its own index, but a recent development has seen the Canadian Forest Fire Danger Rating System (van Wagner 1987) becoming the de facto standard fire danger index (Viegas et al. 1994). The Canadian system is not a single index, but a set of indices with different time-varying characteristics. Its different components are calculated from daily measurements

of air temperature and humidity, wind velocity, and precipitation. The indices can be predicted using forecasted weather variables. In practical terms, its component DC (Drought Code) was originally designed as a numerical rating of the slow-varying moisture of deep, compact organic layers. Over time, it also proved useful to forecast the fuel moisture of live fuels (Viegas et al. 2001). Aguado et al. (2003) showed a good relationship between spatially interpolated DC values and several satellite-derived indices of FMC. The component FWI (or Fire Weather Index—which is a numerical rating of fire intensity) is the most comprehensive indicator of the danger of fire. The European Commission provides an estimation of the fire risk forecast during the peak of the fire season, i.e. from May to October (European Forest Fire Risk Forecasting System or EFFRFS). The EFFRFS has been designed to compute several types of forest fire risk indices varying from short-term (including the Canadian FWI, Figure 19.8) to long-term indices. These long-term indices include the so-called fire probability and fire vulnerability indices. By 2004, this system delivered fire risk maps to a large number of EU Member States and to some Candidate Countries such as Turkey, Bulgaria, and Romania. Examples for 2001, 2002, 2003, and 2004 are shown in Figure 19.8.

The Impact on Ecosystems The most obvious impact of forest fires is the partial or total destruction of the vegetation. This impact has been traditionally measured as the area burnt. This parameter provides a simple estimation of fire impact that can be easily tracked over time and across countries, and it is assumed roughly to relate to both ecosystem and economic impacts. As described above, remote sensing has a lot to offer in burnt area mapping and provides a useful tool that has been combined with GIS techniques to analyse land use dynamics (Lloret et al. 2002), fuel distribution (Koutsias and Karteris 2003), or post-fire erosion hazard (Ruiz-Gallardo et al. 2004). The impact of fire on ecosystems depends on the characteristics of fire, especially its intensity, and this parameter shows great spatial variability, even within the edges of a single fire (Figure 19.9). In high-intensity fires most of the vegetation canopy is destroyed. However, the ability of Mediterranean vegetation to regenerate after fire is well known and this is due to several factors (Buhk et al. 2007; Chapter 7). Some species have structures that protect buds from extreme temperatures such as the bark of the cork oak or of Pinus pinea, and the leaf sheaths of some bunch grasses. Most commonly, many shrubs

Wildfires

2001

551

2002

Fire risk:

Very low

Low

2003

Moderate

High

Very high

2004

Fig. 19.8. Fire risk in Europe according to the Canadian FWI (Fire Weather Index) on 1 July of the indicated years—see text for explanation. Data are from the European Forest Fire Risk Forecasting System at the European Institute for Environment and Sustainability, CEC, JRC, Ispra, Italy; , accessed 5 December 2008.

(e.g. Arbutus unedo, Erica sp. pl.) are able to resprout from below-ground structures that are protected from the high temperatures by the soil. The soil also protects the seeds of some species (e.g. Cistus sp. pl.) that are able to germinate successfully and establish after fire. In

fact, fire itself releases a range of resources (e.g. light and nutrients) that favour vegetation growth. The germination of some species is even increased by fire as it helps to break the hard exterior of their seeds or to open fruits, for example Aleppo pine cones. Site productivity, largely

552

Francisco Lloret, Josep Piñol, and Marc Castellnou

Fig. 19.9. Burned forest in a highly populated area of Catalonia in north-east Spain. Note the spatial variability in fire intensity as indicated by the presence of areas with burned tree trunks and sites where the fire barely scorched the canopies (photo: Marc Santandreu).

due to water and nutrient availability, enhances postfire regeneration so that north-facing aspects, terraces, and valley bottoms, for example, show a greater development of vegetation (Pausas et al. 1999). On the other hand, low-intensity fires reduce the fuel load thereby decreasing the likelihood of future high-intensity fires. Their effects are not so dramatic and they allow the persistence of most tree populations and the soil seed bank. The ability of Mediterranean plant communities to regenerate after fire is not unlimited and high fire intensities reduce the ability of resprouting species to regenerate (Lloret and López-Soria 1993). Although high temperatures increase the germination rate of some species, when they achieve a threshold, seed viability declines (Herranz et al. 1999). As a result, the overall recovery of the community plant cover is negatively related to fire intensity (Díaz-Delgado et al. 2003). The season of fire occurrence may also determine the impact on the vegetation. If fire occurs before the replenishment of the seed soil bank, which for many species occurs in late spring or summer, germination would be scarce the following autumn (Dominguez et al. 2002). Resprouting may also be diminished if fire occurs when the level of stored resources in underground organs is low (winter and spring) (Cruz and Moreno 2001). Repeated fires at short intervals also strongly reduce post-fire regeneration. Seed banks cannot be restored if fire occurs before plants attain their reproductive age. As a result, pine populations may dramatically decline (Eugenio and Lloret 2004). Resprouting also decreases after recurrent fires because the bud bank (the set of meristematic tissues allowing the growth of new shoots) is eventually

depleted, and resources from underground organs are lowered before they can be restored from photosynthetic tissues (Canadell and López-Soria 1998). Overall, plant cover needs more time to recover after successive fires (Díaz-Delgado et al. 2002), and, over time, shifts in vegetation structure and composition may result (Pantis and Mardiris 1992). A recent review of the challenges faced by plants as they recolonize post-fire terrains has been provided by Buhk et al. (2007). Fire also affects some of the services that ecosystems provide, such as water resources and soil protection. In August 1990 a severe wildfire affected the Réal Collobrier experimental research basin (Var Department, France), where catchment hydrology had been monitored since 1966. Depending on the sub-basin, the proportion of burnt surface varied from 0 to 85 per cent. This fortuitous fire allowed a detailed study of the effect of fire on the hydrology of the catchments, as it allowed a direct comparison between affected catchments with controls, and the comparison of each catchment with its own pre-fire conditions (Lavabre et al. 1993). Among other findings, this study showed an increase of c.30 per cent in the annual runoff yield and an increase in flood frequency. The ten-year return period flow estimated before the fire was exceeded three times in a year, even though the rainfall events in that particular year were not unusual (Chapter 8). A post-fire increase in runoff can be considered universal (Belillas and Rodà 1993; Scott 1993), and it is simply the consequence of the reduction in evapotranspiration due to the destruction of the vegetation cover by fire. However, changes in hydrograph form and in flood frequency have not always been found following catchment burning. In a study of fire effects in several catchments in the Mediterranean climate zone of South Africa, Scott (1993) listed several conditions as indicators of a marked hydrological response after fire, and hence, of soil erosion: (1) a high level of soil heating, which itself is a function of fuel load, fuel and soil wetness, and weather conditions; (2) the loss of ground cover may limit vegetation growth and cause the site to be bare for longer after fire; (3) the roads and tracks may become extensions of the channel system after fire; (4) vegetation types that lead to the development of water repellent soils improve the chances of a sharp hydrological response to fire (in Scott’s study, this was high for pine plantations and low for the natural fynboss vegetation); (5) wildfires are expected to cause a stronger hydrological response and a higher erosion risk than less intense fires because they are more likely to occur in hot and dry conditions when soil and fuel moisture are low. Other studies in Italy and Spain have confirmed an

Wildfires

increase in soil erosion at the plot level after fires, and that the increase was more noticeable when the fire was more severe (Giovannini et al. 2001; de Luis et al. 2003). Recent research has also examined the impact of fires on air quality and on the flux of mercury from forest fires to the atmosphere. Cinnirella et al. (2008) used remotely sensed satellite data for 2006 to estimate that the total amount of mercury released to the atmosphere from this source in Mediterranean countries was 4.3 Mg year−1 .

Management: How can the Impact of Wildfires be Reduced? It has been shown that the area burnt increases with the number of ignitions, with the accumulation of fuel, and during adverse weather, and decreases with fire suppression (Figure 19.6a). However, there are interactions among these factors that complicate such a straightforward interpretation. A reduction in the number of ignitions or an increase in the suppression capacity would allow more fuel build-up that, in turn, will positively affect the area burnt (Figure 19.6b). As we have no control on the weather, there are, in principle, three possible ways to fight wildfires: 1. by reducing ignitions; 2. by increasing fire suppression; 3. by reducing fuel loads. The dominant fire-management policy in Mediterranean Europe is the suppression of all wildfires. This strategy, however, has not been able to avoid recent catastrophic fires (Figure 19.4). It has even been argued that it is the extinction of fires that actually causes catastrophic fires (Minnich 1983, 2001). This controversial hypothesis has been called the paradox of extinction. Its rationale is that the systematic extinction of all wildfires allows a build-up of fuel that will be consumed in future large fires caused by very adverse weather. Without fire suppression, there are frequent and small fires, while fire suppression leads to fewer, larger fires. According to this hypothesis, modern large fires arise as an artefact of fire suppression, and they can be prevented by the creation of a landscape mosaic of patches of different ages. The paradox of extinction has generated a strong debate in California in particular. Other researchers argue that the probability of having large fires would be independent of the fire suppression efforts and would respond mainly to extreme weather situations (Keeley et al. 1999; Keeley and Fotheringham 2001; Moritz et al. 2004). In most countries the relevant authorities encourage people to be careful (Figure 19.10) and not to ignite fires

553

in forested areas, and they prosecute those that ignore this advice. Again, the success of this policy is open to question (Figure 19.2) as a reduction in the number of ignitions would obviously reduce the number of fires, but not necessarily the area burnt or the number of large fires in the longer term. The only way left to managers to fight wildfires is by reducing the amount of accumulated fuel, by clearing, burning, or grazing. Grazing and wood collection successfully explain the difference in the area burnt between Mediterranean southern Europe and North Africa (Figures 19.1 and 19.2). The decline of these traditional practices also explains the sudden increase in the area burnt in Mediterranean countries of Europe that occurred in the 1950s and 1960s with the abandonment of rural areas. The example of Spain in this respect is shown in Figure 19.2. However, people will not go back to the countryside, at least to work there, in the same way that their parents or grandparents did. Fuel stocks will stay in place until they are thinned or until the next fire occurs. This fire can be a prescribed fire (a low-intensity fire that is deliberately ignited by skilled teams and burns under controlled conditions), a managed fire (a wildfire that is kept burning freely as long as it is considered to be under control), or a wildfire. This problem is not exclusive to Mediterraneantype areas, but affects many other regions in the developed world. Much of the western USA, for example, has an overwhelming excess of forest fuels (Agee and Skinner 2005). In practice it is very difficult to isolate the effect of each process, namely the number of ignitions, fuel buildup, fire suppression, and meteorological conditions. An experimental approach does not help when large areas and long periods of time are involved, as is the case when dealing with fire regimes. Thus, support for the prevalence of the various component processes usually comes from indirect observations. A comparison of areas with and without active fire suppression from southern California (USA) and northern Baja, California (Mexico) seemed to provide support for the dominant role of fuel build-up (Minnich 1983; Minnich and Chou 1997). However, charcoal accumulation in sediments in Santa Barbara basin (Mensing et al. 1999), and the analysis of a Fire History Database (Keeley et al. 1999) provided support for the assumption that extreme weather events are the key drivers of large fires in Mediterranean California. However, these conclusions can be criticized for several reasons. First, charcoal accumulation may reflect a runoff cycle rather than a fire cycle (Minnich 2001). Second, fire-suppression policy is not the only difference between southern California and Baja California

554

Francisco Lloret, Josep Piñol, and Marc Castellnou

Fig. 19.10. A fire prevention sign in the uplands of Majorca near Luc. The sign is in English for the benefit of tourists (photo: Jamie Woodward).

(Keeley et al. 1999). Third, during the twentieth century other variables have changed (e.g. land fragmentation and climate) besides the increase in the suppression effort. These questions can be addressed using simple models of vegetation and fire spread over homogeneous areas. Such models incorporate meteorological variability, different rates of fuel accumulation, number of ignitions per year, fire-fighting capacity, and prescribed burning. One such model shown in Figure 19.11 was calibrated against fire regime data (number of fires and area burnt per year and distribution among size classes) of northeast Spain, central Portugal, France, and California (Piñol et al. 2005). According to this model, the total burnt area is more or less the same despite any effort to reduce it by extinguishing fires or by using prescribed burning. Nevertheless, the effect of the fire exclusion policy slightly enhances the dominance of large fires, whereas the use of prescribed fires greatly reduces the importance of large fires. In another modelling exercise a similar response was found by varying the number of ignitions—fewer ignitions only slightly reduced the area burnt, but this tends to increase the dominance of large

fires. This approach has been developed further by Piñol et al. (2007) to model fire regime characteristics in a variety of Mediterranean contexts and to assess the degree of uncertainty in the model predictions. Taking into account that large fires are, by far, the most destructive to nature, property, and human lives, it would be more efficient to increase the low rate at which prescribed burnings are conducted in Mediterranean Europe than to invest more resources into increasing the already considerable fire-fighting capacity. Prescribed fires are routinely applied in France (6,000 to 10,000 ha per year) to pastures and to forest understorey. Experimental programmes of prescribed fires are occasionally set in Italy, Portugal, and Spain. In North Africa and the eastern Mediterranean this practice is not carried out and in some countries such as Greece it is against the law to set a prescribed fire. In contrast to other continents, in many Mediterranean countries public opinion does not look favourably on these practices because of the long history of fire being perceived as a catastrophe. In fact, historically, fire has been used for multiple purposes and it is still a useful tool for some management objectives, such as the improvement of habitats and grazing areas

Wildfires

555

(a) 0 Age (years) 50

(b) 300 years

600 years

(c) Fig. 19.11. An illustration of the fire regime model of Piñol et al. (2005) and of its simulation capabilities. (a) Map showing the age of vegetation after 600 years of a simulation calibrated against the fire regime data of Tarragona in north-east Spain. (b) The proportion of the age of the vegetation in the same simulation between years 300 and 600. Vertical black lines indicate the area burnt each year. (c) The proportion of the age of the vegetation between years 300 and 600 in the same simulation but with the proportion of prescribed fire increased to 1 per cent of territory per year. The lightest shading represents recent fires, light grey represents young vegetation, and dark grey represents old vegetation.

or to facilitate the regeneration of some species (Chapters 7 and 23). In this context fuel reduction is the main goal that is achieved by fires themselves. Alternatively, managed fires may also contribute to fuel reduction, as prescribed fires do.

The Future In the previous section the importance of fuel accumulation in shaping the fire regime in Mediterranean regions was stressed. However, climate has also changed during the course of the twentieth century in the Mediterranean and it is expected to continue in the same direction during the twenty-first century in accord with the dynamics of global climate change (Chapter 3). Recent studies suggest that, if the current climatic trends remain constant, fuel conditions in summer will become drier each year and the risk of large fires will increase (Piñol et al. 1998; Pausas 2004). Future climatic scenarios based on global circulation models also predict warmer summers and winters and increased potential evapotranspiration in the Mediterranean basin (Palutikof and Wigley 1996; IPCC 2001; Chapter 3). Although there is no clear agreement on future precipi-

tation trends, it is expected that there will be an increase in rainfall variability (Summer et al. 2003) and, particularly, in the frequency of extreme drought episodes (IPCC 2001). Thus, it is very likely that the climatic firerisk will increase in the near future in the Mediterranean (Mouillot et al. 2002). Whether this increased fire risk will result in a change in the fire regime of the region is another question, as there are many complex interactions between climate, land use, management policies, and vegetation as this chapter has shown. However, the increase in the wildland-urban interface and the further abandonment of marginal agricultural land all points to greater fire activity in the Mediterranean region. Depending on the magnitude of this increased fire activity, changes in the dominant vegetation types may occur, as some species will not be able to replenish their seed banks or resprout with the same vigour after successive fires. Díaz-Delgado et al. (2002) used satellite imagery to show that the recovery of vegetation (measured in terms of NDVI) slowed down after successive fire events. This effect would act as a negative feedback on fire activity, as new fuels will become available more slowly than before. In summary, the prospects are rather gloomy for the Mediterranean region in this respect. In areas where

556

Francisco Lloret, Josep Piñol, and Marc Castellnou

the vegetation is able to cope with the likely increase in meteorological fire risk, then more intense fires will be likely to occur. In areas where vegetation recovery is reduced, the expectation is a reduction in the structure and biomass of the vegetation as well as changes in the dominant species (Chapter 23). These shifts may be enhanced by changes in land use promoted by rural communities as a result of fire, as observed in central Catalonia, where burnt Pinus nigra forests do not regenerate and are replaced by grassland and deciduous forests (Espelta et al. 2002). An alternative is that a sound fire management policy would be able to allow more fires to occur at low intensity (prescribed and managed fires), instead of the destructive high-intensity wildfires waiting in the near future. Fire is a complex phenomenon and its management cannot rely on simple premises. Several strategies must coexist according to specific goals, spatial and temporal scales, and regional idiosyncrasies. Thus, fire-fighting is the first priority when lives and state properties are in danger, but it will not prevent future fires. Instead it will increase the fuel load that will eventually increase the likelihood of catastrophic fires. Fuel reduction by means of prescribed fires cannot be applied to the whole of the Mediterranean region, but it may be an efficient way to keep fuel loads low at particularly sensitive sites such as critical areas for fire propagation. Extensive fuel removal by herbivores or clearing practices may also contribute to reductions in fire extension and intensity. Avoiding non-native flammable species in forests would also help. It is clear that the simplistic view that considers fire as a disaster that can be defeated by improving fire-fighting techniques and by reducing the number of fire ignitions is not realistic. Instead the ultimate causes of the whole phenomenon are structural and due to drivers that determine the outcome of each single event, namely: climate change, vegetation succession, and social changes.

Acknowledgements This chapter is dedicated to the memory of people that lost their life fighting wildfires. We thank Marta Miralles for assistance in collecting statistical data. The research carried out by the authors and reported in this chapter was partially funded by the Spanish MCYT project REN 2003-07198, and the European Project EUFIRELAB. We thank Jamie Woodward and the external referee for reviewing our chapter. The Fire-fighting Corps of the Department d’Interior of the Generalitat de Catalunya also provided support in the development of this chapter.

References Agee, J. K. and Skinner, C. N. (2005), Basic principles of forest fuel reduction treatments. Forest Ecology Management 211: 83–96. Aguado, I., Chuvieco, E., Martin, P., Salas, J. (2003), Assessment of forest fire danger conditions in southern Spain from NOAA images and meteorological indices. International Journal of Remote Sensing 24: 1653–68. Araque, E. (coord.) (1999), Incendios históricos. Una aproximación multidisciplinar, Universidad Internacional de Andalucía, Baeza. Atahan, P., Dodson J. R., and Itzestein-Davey, F. (2004), A fine-resolution Pliocene pollen and charcoal record from Yallalie, south-western Australia. Journal of Biogeography 31: 199–205. Belillas, C. M. and Rodà, F. (1993), The effects of fire on water quality, dissolved nutrient losses and the export of particulate matter from dry heathland catchments. Journal of Hydrology 150: 1–17. Buhk, C., Meyn, A., and Jentsch, A. (2007), The challenge of plant regeneration after fire in the Mediterranean Basin: scientific gaps in our knowledge on plant strategies and evolution of traits. Plant Ecology 192: 1–19. Canadell, J. and López-Soria, L. (1998) Lignotuber reserves support regrowth following clipping of two Mediterranean shrubs. Functional Ecology 12: 31–8. Carcaillet, C., Almquist, H., Asnong, H., Bradshaw, R. H. W., Carrión, J. S., Gaillard, M. J., Gajewski, K., Haas, J. N., Haberle, S. G., Hadorni, P., Müller, S. D., Richard, P. J. H., Richoz, I., Rösch, M., Sánchez Goñi, M. F., von Steding, H., Stevenson, A. C., Talon, B., Tardy, C., Tinner, W., Tryterud, E., Wick, L., and Willis, K. J. (2002), Holocene biomass burning and global dynamics of the carbon cycle. Chemosphere 49: 845–63. Carrión, J. S. (2002), Patterns and processes of late Quaternary environmental change in a montane region of southwestern Europe. Quaternary Science Reviews 21: 2047–66. Sánchez-Gómez, P., Mota, J. F., Yll, R., and Chain, C. (2003), Holocene vegetation dynamics, fire and grazing in the Sierra de Gador, southern Spain. The Holocene 1: 839–49. Castellnou, M., Martinez, E., Pellisa, O., and Reverte, J. (2002), Landscape fire ecology of Pinus nigra, in D. X. Viegas (ed.), Forest Fire Research and Wildland Fire Safety, Proceedings of the IV International Conference on Forest Fire Research: 2002 Wildland Fire Safety Summit, Luso, Coimbra, Portugal, 18–23 November 2002. Chuvieco, E., Salas, F. J., Carvacho, L., and Rodríguez-Silva, F. (1999a ), Integrated fire risk mapping, in E. Chuvieco (ed.), Remote Sensing of Large Wildfires in the European Mediterranean Basin. Springer, Berlin, 61–84. Deshayes, M., Stach, N., Cocero, D., and Riaño, D. (1999b), Short-term fire risk: foliage moisture content estimation from satellite data, in E. Chuvieco (ed.), Remote Sensing of Large Wildfires in the European Mediterranean Basin. Springer, Berlin, 17–38. Martin, M. P., and Palacios, A. (2002), Assessment of different spectral indices in the red–near-infrared spectral domain for burned land discrimination. International Journal of Remote Sensing 23: 5103–10. Cinnirella, S., Pirrone, N., Allegrini, A., and Guglietta, D. (2008), Modeling mercury emissions from forest fires in the Mediterranean region. Environmental Fluid Mechanics 8: 129–45.

Wildfires Cruz, A. and Moreno, J. M. (2001), Seasonal course of total nonstructural carbohydrates in the lignotuberous Mediterraneantype shrub Erica australis. Oecologia 128: 343–350. De Luis, M., González-Hidalgo, J. C., and Raventós, J. (2003). Effects of fire and torrential rainfall on erosion in a Mediterranean gorse community. Land Degradation and Development 14: 203–13. Díaz-Delgado, R., Lloret, F., and Pons, F. X. (2002), Satellite evidence of decreasing resilience in Mediterranean plant communities after recurrent wildfires. Ecology 83: 2293–303. (2003), Influence of fire severity on plant regeneration by means of remote sensing imagery. International Journal of Remote Sensing 24: 1751–63. (2004a ) Spatial patterns of fire occurrence in Catalonia (NE Spain, 1975–1998). Landscape Ecology 19: 731–45. (2004b) Statistical analysis of fire frequency models for Catalonia (NE Spain, 1975–1998) based on fire scar maps from Landsat MSS data. International Journal of Wildland Fire 13: 89–99. Dodson, J. R. (2001), Holocene vegetation change in the mediterranean-type climate regions of Australia. The Holocene 11: 673–80. Dominguez, L., Calvo, L., and Luis, E. (2002), The impact of wildfire season on regeneration of Quercus pyrenaica forest and Pinus sp. stands. Journal of Mediterranean Ecology 3: 47–54. Espelta, J. M., Rodrigo, A., Habrouk, A., Meghelli, N., Ordoñez, J. L., and Retana, J. (2002), Land use changes, natural regeneration patterns, and restoration practices after a large wildfire in NE Spain: challenges for fire ecology and landscape restoration, in L. Trabaud and R. Prodon (eds.), Fire and Biological Processes. Backhuys, Leiden, 315–24. Eugenio, M. and Lloret, F. (2004), Fire recurrence effects on the structure and composition of Mediterranean Pinus halepensis communities in Catalonia (northeast Iberian Peninsula). Ecoscience 11: 446–54. European Commission (2004), P. Barbosa, J. San-Miguel Ayanz, A. Camia, M. Gimeno, G. Libertà, and G. Schmuck, Assessment of Fire Damages in the EU Mediterranean Countries during the 2003 Forest Fire Campaign. SPI 04.64 European Communities; , accessed 5 December 2008. FAO (2001), Global Forest Fire Assessment 1900–2000. Forest Resources Assessment Programe Working Paper 55, Rome. Franco-Múgica, N., García-Anton, M., Maldonado-Ruiz, J., Morla-Juaristi, C., and Sainz-Ollero, H. (2005), Ancient pine forest on inland dunes in the Spanish northern meseta. Quaternary Research 63: 1–14. Giovannini, G., Vallejo, R., Lucchesi, S., Bautista, S., Ciompi, S., and Llovet, J. (2001), Effects of land use and eventual fire on soil erodibility in dry Mediterranean conditions. Forest Ecology and Management 147: 15–23. Grove, A. T. and Rackham, O. (2001), The Nature of Mediterranean Europe. An Ecological History. Yale University Press, New Haven. Herranz, J. M., Ferrandis, P., and Martínez-Sánchez, J. J. (1999), Influence of heat on seed germination of nine woody Cistaceae species. International Journal of Wildland Fire 9: 173–82. IPCC (2001), Climate Change 2001: The Scientific Basis. Contribution of Working Group I, in J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, X. Dai, K. Maskellm, and C. A. Johnson (eds.), Third Assessment Report of

557

the Intergovernmental Panel on Climate Change. Cambridge University Press, Cambridge. Jalut, G., Esteban-Amat, A., Bonnet, L., Gauquelin, T., and Fontugne, M. (2000), Holocene climatic changes in the Western Mediterranean, from south-east France to south-east Spain. Palaeogeography, Palaeoclimatology, Palaeocology 160: 255–90. Keeley, J. E. (1992), A Californian’s view of fynboss, in R. Cowling (ed.), The Ecology of Fynboss. Nutrients, Fire and Diversity. Oxford University Press, Cape Town, 372–88. and Fotheringham, C. J. (2001), Historic fire regime in southern California shrublands. Conservation Biology 15: 1536–48. and Morais, M. (1999), Re-examining fire suppression impacts on brushland fire regimes. Science 284: 1829–32. Kemp, E. M. (1981), Pre-Quaternary fire in Australia, in A. M. Gill, R. H. Groves, and I. R. Noble (eds.), Fire and the Australian Biota. Australia Academy of Sciences, Canberra, 3–21. Koutsias, N. and Karteris, M. (2003), Classification analysis of vegetation for delineating forest fire fuel complexes in a Mediterranean test using satellite remote sensing and GIS. International Journal of Remote Sensing 24: 3093–104. Lavabre, J., Sempere-Torres, D., and Cernesson, F. (1993), Changes in the hydrological response of a small Mediterranean basin a year after a wildfire. Journal of Hydrology 142: 273–99. Lloret, F. and López-Soria, L. (1993), Resprouting of Erica multiflora after experimental fire treatments. Journal of Vegetation Science 4: 367–74. and Marí, G. (2001), A comparison of the medieval and the current fire regimes in managed pine forests from NE Spain. Forest Ecology and Management 141: 155–63. Calvo, E., Pons, X., and Díaz-Delgado, R. (2002), Wildfires and landscape patterns in the Eastern Iberian Peninsula. Landscape Ecology 17: 745–59. MacDonald, G. M., Larsen, C. P. S., Szeicz, J. M., and Moser, K. A. (1991), The reconstruction of boreal forest fire history from lake sediments: a comparison of charcoal, pollen, sedimentological, and geochemical indices. Quaternary Science Reviews 10: 53–71. Mensing, S.A., Michaelsen, J., and Byrne, R. (1999), A 560year record of Santa Ana fires reconstructed from charcoal deposited in the Santa Barbara basin, California. Quaternary Research 51: 295–305. Minnich, R. A. (1983), Fire mosaics in southern California and northern Baja California. Science 219: 1287–94. (2001), An integrated model of two fire regimes. Conservation Biology 15: 1549–53. and Chou, Y. H. (1997), Wildland fire patch dynamics in the chaparral of southern California and northern Baja California. International Journal of Wildland Fire 7: 221–48. Moreira, F., Rego, C. F., and Ferreira, P. G. (2001), Temporal (1958–1995) pattern of change in a cultural landscape of northwestern Portugal: implications for fire occurrence. Landscape Ecology 16: 557–67. Moreno, J. M., Vázquez, A., and Vélez, R. (1998), Recent history of forest fires in Spain, in J. M. Moreno (ed.), Large Forest Fires. Backhuys, Leiden, 159–85. Moritz, M. A., Keeley, J. E., Johnson, E. A., and Schaffner, A. (2004), Testing a basic assumption of shrubland fire management: how important is fuel age? Frontiers in Ecology Environment 2: 67–72.

558

Francisco Lloret, Josep Piñol, and Marc Castellnou

Mouillot, F., Rambal, S., and Joffre, R. (2002), Simulating climate change impacts on fire frequency and vegetation dynamics in a Mediterranean-type ecosystem. Global Change Biology 8: 423–37. Palutikof, J. P. and Wigley, T. M. L. (1996), Developing climate change scenarios for the Mediterranean region, in L. Jeftic, S. Keckes, and J. C. Pernetta (eds.), Climate Change and the Mediterranean. Arnold, London, 27–56. Pantis, J. D. and Mardiris, T. A. (1992), The effect of grazing and fire on degradation processes of mediterranean ecosystems. Israel Journal of Botany 41: 233–42. Pausas, J. G. (2004), Changes in fire and climate in the Eastern Iberian Peninsula (Mediterranean Basin). Climatic Change 63: 337–50. Carbó, E., Caturla, R. N., Gil, J. M., and Vallejo, R. (1999), Post-fire regeneration patterns in the eastern Iberian Peninsula. Acta Oecologica 20: 499–508. Pereira, J. M. C., Sa, A. C. L., Sousa, A. M. O., Silva, J. M. N., Santos, T. N., and Carreiras, J. M. (1999), Spectral characterisation and discrimination of burnt areas, in E. Chuvieco (ed.), Remote Sensing of Large Wildfires in the European Mediterranean Basin. Springer, Berlin, 123–38. Piñol, J., Terradas, J., and Lloret, F. (1998), Climate warming, wildfire hazard, and wildfire occurrence in coastal eastern Spain. Climatic Change 38: 345–37. Beven, K., and Viegas, D. X. (2005), Modelling the effect of fire-exclusion and prescribed fire on wildfire size in Mediterranean ecosystems. Ecological Modelling 183: 397–409. Castellnou, M., and Beven, K. J. (2007), Conditioning uncertainty in ecological models: assessing the impact of fire management strategies. Ecological Modelling 207: 34–44. Quilès, D., Rohr, V., Joly, K., Lhuillier, S., Ogereau, P., Martin, A., Bazile, F., and Vernet, J. L. (2002), Les feux préhistoriques holocènes en montagne sub-méditerranéenne: premiers résultats sur le Causse Méjean (Lozère, France). Comptes Rendues Palevolution 1: 59–65. Riaño, D., Chuvieco, E., Salas, J., Palacios-Orueta, A., and Bastarrika, A. (2002), Generation of fuel type maps from Landsat TM images and ancillary data in Mediterranean ecosystems. Canadian Journal of Forestry Research 32: 1301–15. Ricotta, C., Arianoutsou, M., Díaz-Delgado, R., Duguy, B., Lloret, F., Maroudi, E., Mazzoleni, S., Moreno, J. M., Rambal, S., Vallejo, R., and Vázquez, A. (2001), Self-organized criticality of wildfires ecologically revisited. Ecological Modelling 141: 307–11. Riera-Mora, S. and Esteban-Amat, A. (1994), Vegetation history and human activity during the last 6000 years on the central Catalan coast (northeastern Iberian Peninsula). Vegetation History Archaeobotany 3: 7–23. Rothermel, R. C. (1983), How to predict the spread and intensity of forest and range fires. USDA Forest Service, National Wildlife Coordinating Group, General Technical Report INT-143. Ruiz-Gallardo, J. R., Castano, S., and Calera, A. (2004), Application of remote sensing and GIS to locate priority intervention areas after wildland fires in Mediterranean systems: a case study from south-eastern Spain. International Journal of Wildland Fire 13: 241–52.

Sadori, L. and Giardini, M. (2007), Charcoal analysis, a method to study vegetation and the climate of the Holocene: The case of Lago di Pergusa (Sicily, Italy), Geobios 40: 173–80. Salvador, R., Valeriani, J., Pons, X., and Díaz-Delgado, R. (2000), A semiautomatic methodology to detect fire scars in shrubs and evergreen forests with Landsat MSS time series. International Journal of Remote Sensing 21: 655–73. Scott, D. F. (1993), The hydrological effects of fire in South African mountain catchments. Journal of Hydrology 150: 409–32. Singh, G., Kershaw, A. P., and Clark, R. (1981), Quaternary vegetation and fire history, in A. M. Gill, R. H. Groves, and I. R. Noble (eds.), Fire and the Australian Biota. Australia. Australia Academy of Sciences, Canberra, 3–21. Summer, G. N., Romero, R., Homar, V., Ramis, C., Alonso, S., and Zorita, E. (2003), An estimate of the effects of climate change on the rainfall of Mediterranean Spain by the late twenty first century. Climate Dynamics 20: 789–805. Tàbara, D., Saurí, D., and Cerdan, R. (2003), Forest fire risk management and public participation in changing socioenvironmental conditions: a case study in a Mediterranean region. Risk Analysis 23: 249–60. Terradas, J., Piñol, J., and Lloret, F. (1998), Risk factors in wildfires along the Mediterranean coast of Iberian Peninsula, in L. Trabaud (ed.), Fire Management and Landscape Ecology. International Association of Wildland Fire, Fairfield, Washington, 297–303. van Wagner, C. E. (1987), Development and Structure of the Canadian Forest Fire Weather Index System. Forestry Technical Report 35. Canadian Forestry Service, Ottawa. van Wilgen, B. W., Bond, W. J., and Richardson, B. M. (1992), Ecosystem Management, in R. Cowling (ed.), The Ecology of Fynboss. Nutrients, Fire and Diversity. Oxford University Press, Cape Town, 345–71. Vázquez, A. and Moreno, J. M. (1998), Patterns of lightning- and people-caused fires in peninsular Spain. International Journal of Wildland Fire 8: 103–15. (2001), Spatial distribution of forest fires in Sierra de Gredos (Central Spain). Forest Ecology and Management 147: 55–65. Vélez, R. (2000), Los incendios forestales en la cuenca mediterránea, in R. Vélez (coord.), La defensa contra incendios forestales. Fundamentos y experiencias. McGraw-Hill, Madrid, 3.1–3.52. Viegas, D. X. (2004), Cercados pelo fogo. Os incêndios florestais em Portugal em 2003 e os accidents mortais com eles relacionados. Minerva, Coimbra. Viegas, M. T., and Ferreira, A. D. (1992), Moisture content of fine forest fuels and fire occurrence in Central Portugal. International Journal of Wildland Fire 2: 6–86. Sol, B., Bovio, G., Nosenzo, A., and Ferreira, A. (1994), Comparative study of various methods of fire danger evaluation in Southern Europe, in Proceedings II International Conference Forest Fire Research. Coimbra, ii. C.05, 571–90. Piñol, J., Viegas, M. T., and Ogaya, R. (2001), Estimating live fine fuels moisture content using meteorologically-based indices. International Journal of Wildland Fire 10: 223–40. Viney, N. R. (1991), A review of fine fuel moisture modelling. International Journal of Wildland Fire 1: 215–34.

This chapter should be cited as follows Lloret, F., Piñol, J., and, Castellnou, M. (2009), Wildfires, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 541–558.

IV

Environmental Issues in the 21st Century

This page intentionally left blank

Editorial Introduction Jamie Woodward

This volume has traced the development of the Mediterranean landscape over very long timescales and has examined modern processes in a wide range of settings. Earlier chapters have explored tectonic processes and the evolution of the topography and biota, the nature and impact of Quaternary climate change, and natural hazards, as well as the increasing role of human activity in shaping geomorphological processes and ecosystems during the course of the postglacial period. A core theme in several chapters is the nature of the relationship between humans and the Mediterranean environment. Over the last one hundred years or so, and especially in the period since the Second World War, this relationship has changed dramatically. Resource exploitation, urban expansion, and rural depopulation have all taken place at unprecedented rates, with major impacts upon the quality of land, water, air, and ecosystems. The final part of this volume examines four key topics of environmental concern; its four chapters explore, respectively, land degradation, water resources, interactions between air quality and the climate system, and biodiversity and conservation. Where possible, it is important to place these issues within an appropriate historical perspective. Many components of the Mediterranean environment have responded in a sensitive way to past environmental changes, but the pressures on land and water resources have never been more intense. Improved monitoring networks and new modelling efforts are needed to predict more effectively the impact of climate and social change on all environmental systems and to help inform policymakers seeking a more sustainable use of the region’s resources. Chapter 20 examines the ecological aspects of land degradation and sets out new ideas on productivity dynamics. It explores some of the interactions between land use change, vegetation dynamics, grazing patterns,

and wildfires. The uneven geography of water resources and water use are highlighted in Chapter 21. Water resource issues have become an increasingly important factor in the geopolitics of the region against a background of climate change uncertainty, rising demand, and a diminishing resource base. Chapter 22 analyses the interactions between climate, air quality, and the water cycle. Summertime air pollution adversely affects human health throughout the basin; the stable and stratified summer atmosphere and intense solar radiation—in combination with local pollution circulation systems—is described as a ‘cooking vessel of photochemical smog’. Ozone production, aerosol haze, and pollution plumes from Asia and central Europe are major air quality issues. The final chapter in this book examines biodiversity and conservation across the region—it traces the development of ideas and draws together research from a range of terrestrial and aquatic habitats. Biodiversity and habitat quality are closely related to land, water, and air quality. A key feature of Mediterranean ecosystems is the deep history of interaction with humans— especially through the early use of fire as a management tool and the fundamental changes that accompanied domestication and the development of agriculture. In more recent times, unchecked urban development, especially along the coastal fringe, poses a major challenge to the conservation agenda. Distinctive hotspots of biodiversity with many rare and endemic species have been identified within the basin. These are partly a product of high topographic variability and isolation and illustrate very clearly the fundamental and long-standing links between geodiversity and biodiversity in the Mediterranean world. The protection of these landscapes and habitats is a key challenge for the twenty-first century.

This page intentionally left blank

20

Land Degradation John Thornes

Introduction ‘Land degradation’ means the reduction and loss of the biological or economic productivity caused by land use change or by a physical process or a combination of the two. ‘Land’ means the terrestrial bio-productive system that comprises soil, vegetation, and other biota and the ecological and hydrological processes that operate within the system (UNEP 1992). The main components of land degradation are ecological degradation, soil loss, and reduction in the amount and quality of the available water resources for human survival and economic sustainability. Conacher and Sala (1998) have edited a major volume on land degradation in Mediterranean environments of the world and soil erosion mechanisms and water resources are considered in other chapters of this book (Chapters 6 and 21). This chapter will focus on the ecological aspects of land degradation by exploring some of the interactions between land use change, vegetation dynamics, grazing patterns, and wildfires. This chapter will also try to identify and avoid repeating the myths that abound in the more popular and/or politically motivated accounts of Mediterranean land degradation.

A Brief History of Land Degradation in the Mediterranean Because of the complex spatial mosaic of environmental and cultural conditions across the Mediterranean (see Blondel 2006), it is not simple to identify the causes or main controls of land degradation or the management strategies required to combat degradation (Lesschen et al. 2007; Märker et al. 2008). As discussed in the context of lake sediment records in Chapter 9,

it is certain that the origins of land degradation extend far back into prehistory. Indeed, Naveh and Dan (1973) have proposed a seven-phase history of land degradation for the Mediterranean basin, paraphrased thus: Phase 1 was the Lower Palaeolithic (around 1,000,000 to 100,000 years BP), when the Levant was the main route of biotic and hominid dispersal from Africa to Eurasia and later westwards through the Mediterranean basin. Hunting and gathering were the main activities and the populations were probably very low. Human impact on the environment is not known—but land degradation is assumed to have been negligible. After this, in Phase 2, it is argued that the use of fire as a tool for the opening up of dense forest spread westwards from Greece, possibly reaching France as early as 400,000 BP. Most authors agree that the Neolithic period (Phase 3) was a major time of landscape change. This was the period of domestication of plants, animals, and domestic livestock (Blondel 2006), that again progressed from east to west, being 4,000 years earlier in the eastern than in the western part of the basin (Chapters 5 and 9). During this period, human population increased more than tenfold—a rate of increase not seen again until the twentieth century. Around 4,000 BP, the climate became more arid in southern France and the breeding of sheep and goats was definitely established (Le Houérou 1981). Moreover, transhumance between the garrigue of Languedoc and the upland pasture of the Cévennes was already in progress. Phase 4, from 5,000 BP to the end of the Roman period, saw the domestication of fruit trees, land clearances, and extensive terracing on hillslopes (Chapter 7). Soil and water conservation methods were implemented, for example in Tunisia and in Israel (Lavee et al. 1997). Populations increased and the cutting down of trees for shipbuilding began. In Phase

564

John Thornes

5 the Moslem conquest was, according to Naveh and Dan, accompanied by economic decline in the Levant as the pastoral nomadism of the Arab tribes replaced the developed hillsides and irrigation systems. By contrast, in the Moorish kingdoms of Spain, irrigation was more thoroughly planned and land conservation notably improved by the Arab invaders. Phase 6 of Naveh and Dan is the technological phase from the late nineteenth century, with drainage of swamps, land reclamation, monoculture in agriculture and forest, and the development of large-scale irrigation schemes across the Mediterranean basin. This culminates in a Phase 7 in the twentieth century with the mechanization of agriculture and its industrialization. Le Houérou (1981) disputes the concept that the pastoral and Bedouin civilizations had a greater effect on vegetation and land degradation and believes that the Hellenistic–Roman period had a more drastic impact due to the large populations. Between classical and recent historical periods, Le Houérou argues for a relative stabilization of Mediterranean vegetation, thanks to the slowing down of demographic growth (world population barely doubled from AD 640 to 1650). In Palestine and North Africa, he argues, population actually fell. According to Despois (1961), it seems certain that North Africa has never, in the course of its history, seen as great an extension of crop lands, nor such an increase in population as at the present time. Le Houérou (1981: 481–2) believes that ‘[i]ts vegetation and soils have never been at such a risk’ and that ‘[e]rosion and desertification progress at a terrifying rate. The degradation is essentially the result of human activity; climate does no more than provide favourable, though constant conditions.’ Of course there is no reason to expect a uniform history of degradation across the entire basin. As our knowledge improves in local areas, more complicated histories can be expected to emerge. An excellent example is the study of the Marmora plateau to the northeast of Rabat, Morocco, by Nafaa and Watfeh (2002). They describe the past environmental changes in what is known as the Marmora Forest. The plateau is sprinkled with a forest of cork oak, which used to be one of the largest forests in Morocco, but is gradually being eliminated. Today 180,000 sheep and 6,000 cattle spend more than 300 days per year in the forest, which is considered to be overgrazed. Before 1913, 100,000 trees had been felled. But the degradation of the Marmora Forest cannot entirely be attributed to contemporary exploitation. Studies of the stratigraphy at the coast have shown that, in the Soltanian period (c.22,000– 26,000 BP), the sands overlying the plateau were

reworked extensively by wind action under drought conditions. This is taken to imply that the forest was not in existence at that time, but it appears to have regenerated into a cork oak forest after about 20,000 BP. Moreover, during the Holocene, there is stratigraphic evidence of an alternating sequence of aeolian and colluvial conditions on the slopes, again implying an alternation of steppe and forested conditions (Chapters 4 and 9). As local ecological histories become better known, it seems likely that a series of changes from stable forested conditions to unstable wind-eroded steppe conditions will be revealed. In the last ten years or so, Mediterranean land degradation has come to the political stage for the following reasons: r It has become increasingly obvious that global

r r

r

r

warming, accompanied by regional effects, has become a reality rather than speculation (Chapter 3). The flight from rural areas and the downward spiral of rural services continues unabated after centuries. A series of natural disasters—floods and the drought of the 1980s—have highlighted perceptions of the sensitivities of rural and urban areas to environmental change. The European Union Agenda 2000, seeking to rectify problems with the Common Agricultural Policy, has identified land degradation as a major problem and this has been coupled to a review of soil as a major renewable resource that is impeding the development of a policy for sustainable livelihoods. The EU Water Framework Directive (WFD) (2000) seeks to provide tools for the management and control of river systems using the principles of integrated catchment management. This requires that all river basins are provided with management plans for the monitoring and control of flooding, water quality, and consumable water resources—including groundwater (Chapters 8, 18, and 21). Land degradation within a river basin is important for all of these and therefore necessarily must be part of the monitoring and planning under the Water Framework Directive. Thornes and Rowntree (2006) have expressed some misgivings about the suitability, applicability, and desirability of the unselective application of the WFD across the European Mediterranean countries mainly on the grounds that it is designed for the temperate environments of north-west Europe.

Land Degradation

Contemporary Perceptions of Land Degradation Land degradation in the Mediterranean has been a serious topic since classical times and has been the subject of legislation in all the European Mediterranean countries and in the Maghreb (Burke and Thornes 1998). One example is from Italy, where the most important law for environmental resources, dating from 1989, deals with the reorganization and operation of soil conservation (Freschi 1998). The law defines the planning of activity in a river catchment by means of a basin plan and new central and peripheral organizations are charged with planning and carrying out operations to provide for rational use of water resources and to safeguard buildings and infrastructure from landslides, avalanches, and erosion. A second example is from Morocco, which ratified the United Nations Convention to Combat Desertification (UNCCD) in 1996. Specific measures have been taken to prevent and/or reduce land degradation, to rehabilitate partly degraded land, and to reclaim desertified land (Price 1998). In his report, Price gives a comprehensive account of the threats of, and policy against, land degradation in Morocco, indicating the high level of national awareness of the problem. If further research into local ecological histories reveals a series of shifts from bare ground to a healthy

565

vegetation cover, then a single cycle of the r–K sequence of earlier ecologists might provide a suitable model for the Mediterranean. This approach envisages a change from a clear unvegetated state occupied by exploratory, colonizing plants (r-species) through to a stable cover of mature trees occupied by conservative, more woody plants that are less easily perturbed (K-species). In a recent development of this basic model, Gunderson and Holling (2002) envisage two further phases of collapse (the Ÿ phase) and reorganization (the · phase). In developing this scheme, called ‘panarchy’, they hypothesize the existence of multiple adaptive cycles in time and at different spatial scales as shown in Figure 20.1. The trajectory from r to K replaces the conventional hypothesis of succession and is the logistic curve of modern ecology. The panarchy hypothesis seems to offer a more robust and flexible approach to explaining the history of degradation in the Mediterranean region and has the advantage that its intrinsic dynamics do not need to call upon ubiquitous explanations for growth, collapse, and regeneration of environmental systems that were outlined above. It also avoids the general unified theory approach where a single explanation for degradation is sought for the whole of the Mediterranean region. Because of the increased availability of satellite data gathering and other achievements of modern

Fig. 20.1. Different phases of the Holling and Gunderson adaptive cycle where r is the colonization phase, K is the consolidation phase, Ÿ is the release phase and · is the system reorganization phase. See text for discussion. Modified from Holling and Gunderson (2002).

566

John Thornes

technology, the new data age has given rise to widespread concern over land degradation. This has coincided with the centralization of European bureaucracy and, based on the assertion that land and livelihoods are faced by a severe crisis, has led to a call for a battle against degradation around the Mediterranean. What one might call the ‘Grand European Desertification Project’ started in Mytilene in the Aegean in 1984, with a conference on European Desertification (Fantechi and Margaris 1986). This grew through a series of massive projects in successive research frameworks and is now moving towards a major programme for the mitigation of land degradation across Europe. This is based on the continual assertion of national governments from Greece to Portugal that land degradation is severely undermining planning to stabilize rural environments (Agenda 2000). There is a firmly held belief that the southern countries of Europe are gripped in a crisis of land degradation that can only be exacerbated by the coming effects of global warming (IPCC 2001). Following the UNCOD Nairobi Conference on Desertification in 1977, recognition of the problem in North Africa and the Near East led to ratification of the United Nations Convention on Combating Desertification (UNCCD) and adoption of the associated actions. The problem, as Le Houérou (1981: 482) had previously expressed it in relation to the Maghrebian states, is that ‘Erosion and degradation progress at a terrifying rate.’ There are arguments both against and for this view. In recent years, land degradation has been encompassed by the term ‘desertification’ and thereby has taken on all the uncertainty and political overtones that this term implies (Reynolds and Stafford Smith 2002). Land degradation is a more satisfactory term that avoids many of these ambiguities and difficulties, especially when applied to the Mediterranean. Though ‘desertification’ cannot be regarded as ‘mythical’ (cf. Thomas and Middleton 1994), statements about the intensity and extent of land degradation are at best inaccurate and at worst based on nothing better than guesswork. This is because land degradation in the Mediterranean has been evaluated mainly by soil erosion, which is virtually impossible to measure with any degree of confidence. It is normally expressed as tonnes per hectare and measured on field plots as runoff and sediment yield. The plots are often inadequately designed, poorly located, and operated over too small a time span to produce meaningful results. They are, after all, only a sample in time and space. For this reason, the larger the sample size and the greater the range of conditions covered, the more likely it is that the results will be useful for practical

purposes. So while they often provide valuable experimental results, they hardly provide a basis for policy decisions. One issue in this respect is that it is the loss of productivity that is critical. In the early stages of land degradation—such as the period following land abandonment—because soil nutrients are concentrated in the upper horizons of the soil, organic matter and nutrients are lost in considerable amounts. Although the loss of soil may be small, the loss of productivity can be quite large and land degradation in the wider sense becomes critical. In the search for simple one-number indicators, the amount of soil lost is often advocated as the most valuable indicator of land degradation when that is probably not the case. Perversely, low rates of erosion, when discovered by researchers, may be dismissed as inaccurate because of expectations that Mediterranean rates are much higher than those in temperate environments and the perception of the Mediterranean as a ruined landscape or lost Eden. The reasons for, and history of, this misconception are discussed in Grove and Rackham’s The Nature of Mediterranean Europe: an Ecological History (2001). As well as data from global modelling, empirical data do, in fact, reveal that the Mediterranean is a region of relatively high soil erosion (Chapter 6), even recognizing the very high spatial variability. So far, most estimates of erosion have been based on the USLE (Figure 20.2). The underlying approach has also been used for the estimation of soil losses based on the CORINE (Coordination of Information on the Environment) GIS system. The maps produced and the accompanying table indicate the actual erosion risk in the European Mediterranean (Table 20.1). This seems to indicate that the largest absolute area of high-risk land is in Spain, with about 30 per cent of land in the high risk area (Figure 20.3). The actual soil erosion risk map shows that the risk is TABLE 20.1. National soil erosion risk data for five Mediterranean countries in the EU Country

High risk area km2

%

Moderate risk area km2

France (south) 1,693 1 22,362 Italy 30,169 10 93,983 Greece 27,713 19 47,877 Spain 145,494 29 219,908 Portugal 26,878 30 48,166 EC (south) 228,947 19 432,295

Low risk area

%

km2

%

12 31 36 44 54 36

123,643 165,823 39,287 111,518 12,884 453,155

65 55 30 23 15 37

Excluded area km2

%

42,463 22 11,303 4 20,113 15 20,598 4 1,000 1 95,477 8

Note: The data for Spain and Portugal do not include the islands. The excluded area column comprises urban land, lakes, bare rock terrains, and areas without data.

Land Degradation

567

Fig. 20.2. An application of the USLE for the Autonomous Region of Andalucia, southern Spain. Source: Ministerio de Media Ambiente, Junta de Andalucia. Note the very high rates of erosion in the east of the area near Almeria (Chapter 1).

concentrated in Portugal and southern Spain. All these data must be viewed with caution, though they are the best available estimates.

Land Degradation: The Main Processes The Role of Vegetation Land degradation, as defined above, is absolutely and inextricably linked to the vegetation cover and functional type. The vegetation cover is the percentage of the ground that is covered by vegetation when viewed vertically from above. This directly controls the amount of erosion that occurs in relation to the bare ground value (Figure 20.4; Chapter 6). As the percentage of cover decreases, the amount of erosion increases. This increase in erosion is most rapid just below 30 per cent vegetation cover. Salvucci and Eagleson (1992) have demonstrated that the ecologically optimal cover

in dry lands with this value occurs when the ratio of actual to potential evaporation is also about 30 per cent. Below this, erosion rates increase catastrophically (Elwell and Stocking 1976; Francis and Thornes 1990). Functional type refers to whether the plants are grasses, shrubs, or trees and the functional type depends on the trade-offs that the plants have to make in capturing and storing different resources, notably light and water (Tilman 1982). This means that where water is scarce and light is abundant, more carbon is allocated to roots and stems than to leaves. Mulligan (1998) has shown that these functional types affect erosion rates. Human exploitation of the vegetation cover also depends on the functional types—grazing animals have a strong preference for leaves, firewood gatherers for woody stems and branches. The vegetation evolves in response to the controls of light, water, and temperature to reach a steady state in pattern, amount, and functional type. Erosion responds accordingly. In a stable ecosystem, physical or

568

John Thornes

Percentage of land at high erosion risk 80 – 100 60 – 80 40 – 60 20 – 40 0 – 20

Fig. 20.3. The EU erosion estimate for Spain based on the CORINE database showing the percentage of land at different levels of erosion risk. See text for discussion.

human perturbations have little effect and the system can restore itself to the previous steady state. A simple metaphor (Figure 20.5) is to regard the values of the system state variables as a landscape over which a ball is rolling. The ball comes to rest in the depressions and rolls away from the bumps. The depressions (attractors) are those values of the system variables where stability prevails (the ball stops moving). By contrast, the bumps are unstable locations that the ball moves away from (repellers). If the ball is in a depression energy is needed to move it to another position (changing its state). The deeper the depression, the more energy is needed to change the state and the more stable is the system. On a flat plain the ball rolls about at random (varies randomly in time, but only by tiny amounts). A fuller

description of this metaphor is given in Holling et al. (2002a ). Unstable states give large fluctuations through time in the vegetation cover (and hence in erosion). Generally speaking, droughts provide instability in Mediterranean ecosystems as the vegetation tries to reach ecological optimality with respect to the changing hydrological conditions. Under severe perturbations, the system may start to oscillate widely with large underand over-adjustments. This situation was modelled for dry Mediterranean conditions by Thornes and Brandt (1994) using real rainfall records and a boom-andbust cycle in vegetation cover and erosion was the result. Pueyo et al. (2007) have argued that land use fragmentation and aridity were the key controls on

100

100

75

75

Soil loss

50

50

25

25 Runoff 0

0 0

25 50 75 Mean seasonal vegetation cover (%)

Runoff as a percentage of mean annual rainfall

Soil loss as a percentage of bare plot results

Land Degradation

569

(a)

(b)

100

Fig. 20.4. Soil loss and runoff as a function of the proportion of ground covered by a vegetation canopy (after Elwell and Stocking 1976).

the extent of vegetation recovery between 1957 and 1998 in parts of the middle Ebro valley in north-east Spain. Interannual rainfall variability is greater in dry Mediterranean regions (Chapter 3), therefore the ecosystems there appear to be more unstable. As the climate becomes drier or wetter the interannual variability reduces, as does the magnitude of erosion. After fire in the more humid Mediterranean environments, such as the Catalan countries of Spain and the mountain areas of Greece and Italy, field experiments appear to indicate that after 8–10 years the rate of sediment production typically falls back to levels that existed before the fire. In drier areas, with greater vegetation instability, the impacts of fire should be greater and longer-lasting, especially if the vegetation cover is close to the 30 per cent cover level. (See Chapters 6 and 7.) The overall conclusion is that, where wash erosion dominates, rates should be highest where it is climatically drier, but not arid. This reinforces the opinion that rates of erosion are highest in dry Mediterranean areas with rainfall around 250 mm and with actual evapotranspiration less than 30 per cent of potential evaporation. This matches with Langbein’s and Schumm’s (1958) empirical curve of sediment yield for the conterminous United States that shows a peak of sediment yield at approximately 300 mm per year (Figure 20.6a). Kirkby’s (1980) model-based analysis shows a similar peak for wash erosion (Figure 20.6b). On the semi-arid to arid side of this curve, as runoff decreases, so too does sediment yield. On the wetter side there is a very

(c)

(d)

a b

Metaphor

Phase space

Fig. 20.5. A metaphor for a system’s stable and unstable conditions (modified after Gunderson and Holling 2002); (a) is a level surface, (b) has a slight depression, (c) has a hill, and (d) has two hills and two depressions. See text for discussion.

steep fall to lower suspended sediment yields as the vegetation cover increases and shifts from shrubland to grasses and eventually to forest. Kosmas et al. (2002) have gathered a large sample of erosion plot data that appears to confirm the general behaviour described

570

John Thornes 1,500

(a) 20

1,000 Forest 750

Grassland

500

250 Desert shrub

(c)

15

10

5

0

0

100–200 200–300 300–400 400–500 500–600

re

300 100 200 0 Annual sediment yield (tonnes/km2)

1,000

Ba

Annual erosion from a 10-m long 10° slope (mm)

Mean annual sediment yield (g m-2)

Effective precipitation (mm)

1,250

Annual precipitation (mm)

(b)

rub

Sc

100

Gras

Curves for constant vegetation cover

d

slan

Composite curve for natural vegetation

10

Bare

1

t Fores

0.1

Curves for constant vegetation cover Composite curve for natural vegetation

Water erosion

Wind erosion (very approximate)

Scrub Forest

Grassland

0.01 0

500

1,000

1,500

2,000

Mean annual precipitation (mm)

Fig. 20.6. (a) The relationship between sediment yield and annual effective rainfall (that which produces runoff) from Langbein and Schumm (1958). (b) Estimated rates of erosion by wind and water as a function of rainfall and vegetation cover (Kirkby 1980). (c) Erosion on field plots around the Mediterranean, as a function of rainfall (after Kosmas et al. 2002).

above that, at the global scale, the Mediterranean is a region of relatively high rates of erosion (Figure 20.6c) (Chapter 8).

Sediment Yield Data Given the difficulties of field-plot data referred to above, it is instructive to take note of measured suspended sediment yield data for Mediterranean rivers gathered and interpreted by Woodward (1995) who starts with a quotation from Zachar (1982: 389), ‘The greater the effects of human intervention and the more extreme the conditions of climate, the more serious the consequences of erosion have been. Thus the coun-

tries around the Mediterranean are most affected by erosion.’ The data in Woodward’s analysis appear to contradict the empirical (Langbein and Schumm 1958), theoretical (Kirkby 1980), and simulation (Thornes and Brandt 1994) results described above, but draw attention to other causes of high sediment yield. These are deep gully and badland development, exceptionally erodible lithologies, steep terrain, high-intensity rainfall and flood-prone catchments, and overgrazing by sheep and goats. He also points out that there are technical problems with the use of suspended sediment yield regressed against rainfall (spurious correlation resulting from co-variation) and, following Zachar (1982),

Land Degradation

the bias introduced by locating measurements at sites designed to measure soil loss for agricultural planning. There is, in addition, a bias towards small research basins with their very high sediment delivery ratios. The same difficulties seem to apply to reservoir siltation studies. The developments in Caesium 137 (137 Cs) analysis and stable isotope studies may provide the solution in the search for erosion estimates in Mediterranean environments (Navas and Machin 1991; Schaller et al. 2001; Zorzou 2004). In conclusion it seems that erosion rates are highly variable across space and over time and that, as yet, no reliable widely available method has been produced that enables us to compare the relative degrees of erosion in different countries across the Mediterranean basin. The most we can hope for from direct measurement of the types described is the possibility to confirm or deny the wide range of existing modelling attempts.

Wind Erosion and Salinization Two other often-emphasized degradation processes are wind erosion and salinization of soils. Wind erosion and sand movement are generally minor concerns in most natural frameworks in the European Mediterranean (Warren and Barring 2002; Riksen et al. 2003), except in small areas such as central Aragon (Gomes et al. 2003), coastal Turkey, Sicily, and Sardinia, but they are an important form of land degradation on the Saharan edge of the Maghrebian states, such as Tunisia and Algeria. In Tunisia the main area of aeolian deflation and accumulation in nebkhas and dunes is in the south-west of the country where major conservation efforts are continually carried out (Mtimet 1999). Further discussion of aeolian processes in the Mediterranean region is given in Chapter 14. Salinization is a major problem in Tunisia where saline phreatic water has destroyed 16,000 hectares in the lower Mejerda valley and the oases of Tozaur, Gafsa, Kebili, and Gabes have been significantly affected. Salt intrusion to coastal groundwater occurs where sea water invades aquifers near or at the coast (Chapter 21). The alluvial plain of Capoterra in southern Sardinia has been studied by Barrocou (1994). Salt water intrusion has resulted from over-exploitation of groundwater to satisfy the demands of agriculture, industry, and domestic use. Groundwater is withdrawn from the aquifer system from about 300 wells scattered over an area of nearly 60 km2 . Continuous monitoring has indicated a rise in salinity, particularly in the central part of the alluvial plain.

571

Groundwater Abstraction A related theme across the Mediterranean, both north and south, is the withdrawal of groundwater, notably for irrigation (Chapter 21). In Castilla la Mancha, for example, the area under irrigation increased from 15,916 ha in 1970 to 352,452 ha in 1996. There have also been reduced inputs to the aquifers through declining rainfall, especially in the general decline of the 1960s and 1970s and the dry years of the early 1980s. On top of this, a general rise in temperature has led to increasing evapotranspiration losses (Burke 1998). These hydrological budget changes have led to land degradation. In the wetlands of Tablas de Damiel (Llamas 1988) at the confluence of the Ciguela and Guadiana rivers, 20 km2 of wetland has been reduced to less than 1 km2 as a result of abstractions from the in-flowing rivers and extraction of water from the underlying aquifer. So desiccated have the wetlands become that the residual peat is also disappearing as a result of internal combustion.

Precursors of Degradation In the light of the discussion of vegetation and erosion above, it is axiomatic that changes to the canopy can be expected to lead to higher rates of erosion (Chapters 6 and 8). This is why Mediterranean land degradation is most commonly ascribed to the processes of deforestation, overgrazing, and land abandonment. The converse of deforestation as a cause of land degradation is to propose afforestation as a solution. In fact both deforestation and grazing are very variable and complicated processes. Before and after effects are very difficult to establish and paired catchments are almost impossible to compare. As Chapter 6 has shown, the underlying terrain that is exposed is also very variable as regards soil type, depth of soil, and roughness conditions.

Deforestation There is much more to the story of deforestation than is usually supposed. The meaning of ‘forest’ in historical texts is enormously variable, ranging from Mediterranean scrubland to previously planted trees. The firmest evidence of conditions before deforestation comes from long sequences of pollen records (Chapters 4 and 9). Some of the best sites for pollen preservation are in upland lakes and bog sites that, for climatic reasons, were almost invariably clothed with mature forests of woody trees with dense canopies. In some lowland areas preservation is more difficult and cave sites may poorly reflect the surrounding pollen rain. A few

572

John Thornes

exceptional sites, such as Padul near Granada, Spain, Lago di Monticcio in southern Italy, and Lake Ioannina in north-west Greece, provide very long sequences that have been extrapolated over wide areas (Chapter 4). There are other problems in interpreting prehistoric vegetation from pollen. It can be argued that marine sediments are difficult to interpret because the origin of the pollen grains is often difficult to establish. Also, as Grove and Rackham (2001) note, palynology depends very largely on wind-pollinated plants whereas, in the Mediterranean, whole categories of plants are insect pollinated and leave very little to provide a pollen record. For an extended discussion, the reader is referred to Wainwright and Thornes (2003). Grove and Rackham (2001: 157) summarize the position for Mediterranean Europe thus: By 7000 BC there were trees in all parts of the Mediterranean that have a pollen record. Usually oaks predominated, but in Corsica the predominant tree was tree-heather and in parts of Spain, south Greece and Crete it was pine. Often there was more, sometimes much more, deciduous oak than evergreen. There is general evidence for other typical Mediterranean trees and shrubs, Pistacia (lentisic or terebinth), Phillyrea, Arbutus, plane, olive. There were also northern trees: alder, birch, elm, hazel, hornbeams (carpinus and ostrya), lime (particularly significant because it is insect pollinated).

It can be argued that, given large variations in time and space, it is extremely difficult, on the basis of the limited sample of pollen sites in what are often ecologically unique situations, to make conclusive generalizations about the state of prehistoric and early vegetation cover. However, well-dated long-term pollen-based records of vegetation change from across the Mediterranean are presented and discussed in detail in Chapters 4 and 9. There is also the problem that there have been other causes of forest loss in addition to deforestation by clearfelling. Many other forms of ecological disturbance result in the loss of woodland, such as wildfires and changes in the fallow cycle and the introduction of new animal and plant species. On some of the Greek islands, practices related to grazing and burning have also adversely affected woodland conservation. The woodland cover has been subjected to climate change that has resulted in changes in the ecologically optimal cover. There is also a problem of circularity if erosional and sedimentological changes are used to infer vegetation and climate changes in one area and the climate change is then used to explain the vegetation in another. Even if the picture of deforestation were clear, there would be some reservations about the relationship between forest clearance and degradation. Many studies

of the impact of fire reveal dramatic changes in runoff and sediment yield in the period immediately following the burn. Usually, however, after 8 to 10 years (depending on the weather sequence), the runoff and sediment yield rates have fallen back to the pre-burn levels. Another factor affecting the impact of fire and clearance is the extent to which the organic litter layer is destroyed. Brandt (1989) also showed that, for both temperate and tropical forests, the mean raindrop energy beneath the canopy was higher than that beneath open skies, presumably due to the concentration of water on leaves. Without a protective litter layer, this can yield higher soil loss rates beneath forest. Again the recovery rate and pattern are relevant (Obando 2002). In conclusion, clear felling of trees almost invariably produces increased runoff volumes and enhanced sediment yields. Nevertheless, field experiments on the impacts have proved difficult to evaluate even though they have been conducted for almost 100 years since the celebrated Wagon Wheel Gap paired catchment studies that started in southern Colorado in 1910. Prior to this experiment, it was generally assumed that: forests reduce the magnitude of ordinary seasonal floods; forests maintain stream flow in dry weather; forests prevent erosion. The essential reasons for uncertainty about the experimental outcomes—which, in general, did not support these assumptions—are discussed by Rodda (1976).

Land Abandonment Land abandonment has become a major feature of the European Mediterranean countries over the last century (Lesschen et al. 2007; Chapters 7 and 23). This has been largely attributed to rural depopulation that affected most of Europe at the same time, a mixture of declining birth rates, the failure of rural service provision, and the growth of economic opportunities. In some areas, special circumstances led to rural abandonment. Thus, in the eastern Alpujarras of Andalucia the destruction of the vine stocks by the disease phylloxera at the end of the nineteenth century had an impact on rural populations that the area never recovered from. Later, in the twentieth century, economic conditions led to further out-migration, with migrants seeking employment in large Spanish cities (notably Barcelona), in coastal resorts, and in Germany under the Gast Arbeiten scheme. In the 1980s a major drought also pushed rural population to the brink and led to a further exodus. Much of European Agricultural Policy in the last forty years has been directed to slowing or halting this flight through financial incentives (Thornes 2002).

Land Degradation 50

40

20

10

0

–10 <10

10–25

25–50

>50

Age of abandonment (years) Fig. 20.7. The evolution of plant cover over time on abandoned plots in the Spanish Pyrenees (after Garcia Ruiz et al. 1994).

There have been several studies of the relationship between land abandonment and land degradation. Garcia Ruiz et al. (1994: 28) described farmland abandonment as ‘probably the most important geo-ecological process affecting soil and water in mountain areas of Central and Western Europe’. They concluded that farmland abandonment produced many changes in runoff and sediment yield. Land management after abandonment was the most important factor in preventing any undesirable effects from a geomorphological and hydrological point of view. An important issue was the recovery of the vegetation after abandonment, dependent on the interaction between

25

Runoff coefficient (%)

20

land uses after abandonment (including grazing and fire) and the physico-chemical characteristics of the soils. Figure 20.7 shows their data on the evolution of the bush cover following abandonment. The temporal pattern shows a steady but steepening increase of bushes up to twenty-five to fifty years after abandonment, followed by a levelling off at about 30 per cent plant cover after fifty years. Figure 20.8 shows the results they obtained from plots after a period of eighteen months of observation of runoff and sediment yields. It shows very clearly that shrub cover can be an excellent plant formation for soil protection. The results also support the use of the logistic curve for vegetation recovery after abandonment, as do similar studies after fire (Godron et al. 1981). In the logistic curve there is a slow recovery at the start of recolonization (the r-phase referred to above). With feedback the process gains momentum leading to a steep rise. As resources become scarcer (water, light, and nutrients) a limit to growth sets in and the rate of increase of biomass levels off as the vegetation approaches the potential when the rate of growth falls to zero or varies around the potential. This is the K-phase of the panarchy paradigm (Figure 20.1). Field surveys in the semi-arid Caraco basin in Murcia showed that abandoned fields had more gully erosion than cultivated ones (Lesschen et al. 2007). By simulating patterns of land abandonment with four different land use scenarios for the period 2004–15, these researchers showed that, for the various land use scenarios, the potentially vulnerable areas for gully development increased from 18 to 176 hectares. Caution is needed, however, with respect to studies that show only loss of soil mass. From an agricultural point of view, what matters more is the loss of nutrients, which is a sensitive indicator of land degradation.

1,000

Runoff coefficient Sediment yield

800

15

600

10

400

5

200

0

0 Shrub cover Shrub cover 100% 85%

Meadows 85%

Shrub cover Shrub cover Shrub cover 60% 40% 15%

Sediment yield (g/m2/yr)

Plant cover (%)

30

573

Fig. 20.8. Runoff and sediment yield data for different land uses on regenerated abandoned fields in the Spanish Pyrenees (Garcia-Ruiz et al. 1994).

574

John Thornes

A recent study by Pardini et al. (2004) carried out in the Cap de Creus Peninsula of north-east Spain, emphasized this aspect of land degradation. Small losses of topsoil can result in rapid and large losses in productivity. This fact is not often recognized in the soil erosion league tables for the Mediterranean. As Pardini and co-workers have shown, the most significant outcome is not the absolute loss of soil, but depletion of the organic carbon and total nitrogen stores—especially in burnt environments. They go on to conclude that abandoned soils left under natural vegetation are comparatively less susceptible to erosion and nutrient depletion than reforested or cultivated soils. Wildfires determine in part the sequence of shrubs that occurs (Chapter 7). When there is replacement of ericaceous shrub by Cistus shrub, this tends to create less favourable soil physico-chemical conditions and therefore increases the likelihood of degradation. In this example, forests planted with Pinus halepensis were so detrimental to soil properties that reforested areas showed lower organic content and cation exchange capacities and therefore higher susceptibilities to erosion. The regeneration after clearance by felling or following cultivation is largely affected by the weather sequence and, since agriculturalists do not have much control over the timing of their operations, regeneration is a very hit-and-miss process. Because of the global influence of ENSO (El Niño Southern Oscillation) and the connection between the North Atlantic Oscillation and the Mediterranean (Goodess and Palutikof 2002: Chapter 3), The cyclicity of Mediterranean rainfall has become more predictable. Regeneration planning could respond to these oscillations by taking advantage of runs of relatively wetter years rather than a dry series of years for planting, as observed by Holmgren and Scheffer (2001). The quicker that a vegetation cover greater than 30 per cent can be established, the less will be the resulting land degradation. Controlled experiments are difficult to achieve, but it is often helpful to carry out computer simulations in order to experiment with forest strategies. For example, in an investigation by Obando (2002), based on field studies near Lorca in south-east Spain, simulations on forest regeneration, following land clearance in a dry Mediterranean environment, examined the effects of patterns of abandonment (leading to regeneration) on the sequential characteristics of the recovery. She sought to explore the effects of four variables: different initial conditions; different patterns of abandonment; different rainfall regimes in the regeneration period; and different time-patterns of abandonment (the percentage

of land abandoned per year). She modelled runoff and erosion from a cellular grid overlying a basin topography with soil-covered slopes. The model comprised a coupling of hillslope hydrology, vegetation growth, erosion, and runoff to the outlet of the basin. The model also accommodated spatial patterning selected to allow sections of the catchment that were considered vulnerable to high erosion rates to be abandoned experimentally (e.g. areas with steep gradients or lithological crusts susceptible to erosion). Obando’s experiments produced a sequence of annual vegetation covers in the basin and the corresponding runoffs and sediment yields under the various conditions. The annual runoff generated by the model for the bare catchment is higher than that for the shrubcovered catchment over a period of twenty years. Under dry conditions (R < 100 mm) annual runoff is proportionally lower in both cases. The difference between the two catchments decreases for events >110 mm and the runoff coefficients obtained using the wet series for the bare and shrub-covered catchments are 33.5 and 21.8 per cent respectively. The daily biomass production (g/m2 ) closely tracks the rainfall and, for the bare areas (with an initial seed-bank of 0.1 g/m2 ) quickly (six years) reaches the production rate of the heavily vegetated (330 g/m2 ) plots. Because these are real rainfall series, the rate falls again after about fourteen years. In other words, in this model result, the effects of rainfall fluctuations are more important than the final steadystate capacity vegetation towards which the system is ‘attracted’. The key outcomes of this work may be summarized thus: 1. For the three rainfall series and for the selected types of abandonment, the simulated sediment yield varies with rainfall amount and intensity as well as with the percentage and pattern of abandonment. 2. The sediment yield decreases with increasing vegetation cover following abandonment but, for very high rainfall intensities, the sediment yield for the vegetated catchment is not very different from the bare catchment. 3. In order to reduce the sediment yield, the best strategy is to select the most vulnerable areas for abandonment, while using those less vulnerable for agriculture. 4. The highest sediment yield is produced when areas close to the channel are not abandoned. These results have significant implications for the management of Mediterranean catchments. They show

Land Degradation

that control over the time-space pattern of deforestation/land abandonment can be used to reduce land degradation even at times of climate change and call for a more thoughtful approach to basin management from the forestry industry. They further invite caution about the assertion that deforestation is the main agent of Mediterranean land degradation by demonstrating the wide variety of possible responses.

Grazing Grazing is so frequently blamed for land degradation in the Mediterranean as to merit special attention in this chapter and some examples of blame are given below. There is need for some caution in adopting a singleprocess explanation because of the variable character of the grazing/land degradation interaction in space and time, the relatively poor understanding of the phenomenon, and the tendency of managers and politicians to seek a simple explanation for what appears to be a ‘quick-fix solution’ that can be easily implemented such as a limited and fixed ‘carrying capacity’. There are five main positions in the debate. First there is the view of those who consider that grazing by domestic animals is among the major causes of land degradation, with goats being singled out for their predilection for woody forage (e.g. Thirgood 1981, 1988). Tsouris (1985) claims that the pastoral economy had a much greater impact on deforestation in the Mediterranean than agricultural clearances. Second, there are animal ecologists (e.g. Papanastasis 2004) who claim that, since livestock numbers have been drastically reduced, in some parts of the Mediterranean undergrazing has resulted and this has led to the piling up of flammable biomass and hence increased fire risk (Chapter 19). Third, as with the forestry debate, there were quite different initial conditions in different parts of the Mediterranean, with livestock grazing spreading at different rates in the Holocene, being introduced 3,000 years later in the west than in the east. And they did not find a grazing-free environment, since wild herbivores were already there. In Crete, for example, wild fauna had arrived by the Middle Pleistocene (750,000–128,000 years BP) (Rackham and Moody 1996). Fourth, Papanastasis (1998) makes the point that Mediterranean ecosystems were affected not only by the grazing per se, but by the whole livestock economy, which involved other activities too such as shepherding, nomadism and transhumance, sheltering, milking, cheese-making, and especially burning. The latter (still practised today) was used to create grazing and suppress unwanted, chiefly woody vegetation in range-

575

lands. The fifth approach to the debate attempts to establish the detailed process dynamics of the plant-animal interaction. One example is the work of Briske and NoyMeir (1998) that showed that several ecological scales must be considered when evaluating the response of vegetation to grazing to minimize incomplete and inaccurate conclusions. Ultimately, biomass reduction by consumption and by selective grazing leads to loss of cover and thereby to increased erosional susceptibility, but there are other effects too. The soil can be compacted, leading to crusting and/or higher levels of runoff and erosion. Regular animal routes (pathways) can become concentrations of overland flow that lead to gullying. Concentrations of livestock around watering points can produce the socalled ‘piosphere’ effect, whereby badly degraded lands form at these point locations with vegetation decreasing towards the water points. Some authors argue that the increased addition of manure has the reverse effect. The overall balance appears to depend on cattle density and raises the question of rapid change after a threshold has been crossed. Another important effect is the change in composition of the vegetation caused by the selective grazing of the more palatable plants. Recently it has been shown that erosion can develop and spread from erosion hotspots (Thornes 2005). There are two main processes at work here: the origin of the bare spots amongst otherwise fairly uniform vegetation cover and the spreading out of erosion from these hotspots. The original spots may be random or uniform and circular or linear and are often attributed to cattle concentrating for resources (food or water) or to random variations in topography or soil quality. Hotspots have been reported in the dry north-western sector of the island of Lesvos, Greece, but they have also been reported beyond the Mediterranean too, for example in the sub-humid lands of the Eastern Cape, South Africa (Kakembo 2003) and in the much wetter and cooler Icelandic heathland (Gisladottir 2004). A study of the emergence of such bare spots reveals that they can arise from uniform vegetation covers when there are positive feedbacks in the relationship between moisture, vegetation growth, and animal dispersion. They produce spotty patterns in the vegetation biomass cover that need no special initial patterning to develop and evolve. This approach (Cronhjort 2002) also shows that the phenomenon can lead to apparent movement of island-like patches of vegetation through time and that the emergence is related to climatic variations. A recent study from the European Union GeoRange Project has used the concept of ‘resistance’ surfaces. In this approach, the available forage for animals is

576

John Thornes

mapped from satellite data and the barriers to movement (such as steep rocky slopes or poor grazing) are used to identify the optimal grazing location. This, taken in turn with soil and hillslope hydrological information, facilitates mapping and prediction of erosion-susceptible areas due to grazing. This methodology appears to offer great possibilities for future grazing management. Both approaches operate on representations of the landscape as cellular maps and hence are suitable for the application of cellular automaton modelling techniques that are rapidly being developed in Spatial Ecology (Dieckmann et al. 2002). As the impact of global warming becomes more acute in the Mediterranean, the threat to grazing difficulties seems likely to increase. Indeed, in a study of stocking capacity resilience to changing climate, Köchy et al. (2008) have shown through hierarchical modelling that grazing can result in greater reductions in 3,000

productivity as climate becomes drier. (Thornes and Brandt 1994) showed that the competition between vegetation and erosion leads to an oscillatory behaviour in both that arises from the intrinsic dynamics, prompted by external forcing (Figure 20.9). This is represented in a ‘boom and bust’ cycle. The forage increases dramatically in good years but then is unable to survive in dry years and drops to very low levels, thus leading to severe erosion. This can happen without cyclical rainfall series but it seems likely that global warming might progressively destabilize the grazing/soil erosion interaction and that cyclical behaviour is a symptom of this instability in the system. Ongoing research based on the BIOME model by Zhang at King’s College London (personal communication) has compared the land use of 1981–90 to 2040–50 with respect to climate change, as indicated by accepted European climate models. It shows a marked reduction (15%) of xerophytic

600

(a)

2,000

400

1,500 1,000

200

Rainfall (mm)

Biomass (g m-2)

2,500

Vegetation

500

Rainfall 0

0 0

2,000

20

40

60 Year

80

100

120 600

(b)

Vegetation Rainfall 400

1,000 200

Rainfall (mm)

Biomass (g m-2)

1,500

500

0

0 0

20

40

60 Year

80

100

120

Fig. 20.9. Simulated response of vegetation to stochastic rainfall generated by a rainfall generator model for Murcia, Spain; (a) with favourable conditions but with 10 cm of soil already lost, and (b) with a decreasing rainfall of 3 mm per year (after (Thornes and Brandt 1994).

Land Degradation

wood/scrub and a dramatic reduction (nearly 96%) of cool grass/shrub. The benchmark (1981–90) was validated against Olson’s European remotely sensed land use data. The results should be taken as a warning of real and serious changes ahead. The destabilization of the grazing/soil erosion interactions would push the system into greater instability than that which already exists. In recent years, important lessons have been learned from the grazing management practices of sub-Saharan savanna rangelands, which should be relevant to Mediterranean pastoralism. Scoones (1989) has traced the history of colonial rangeland management practices and Behnke et al. (1993) have identified their weaknesses. The main shortcoming was the rigid application of the notion of carrying capacity and the development of elaborate algorithms produced to estimate it. The main problem is that a single number is being used to represent a complicated, complex, and highly variable interaction (the indicator fallacy). Second, it implies a stable situation in which management is all that is needed to maintain the balance. Third, the approach allows little insurance in times of global climatic or economic instability and pastoral livelihoods are undermined. The preference now is for a more flexible, opportunistic approach that provides the farmer with a wider range of alternative methods of adjustment to extreme factors. A more effective management strategy calls for a better recognition of the thresholds and instabilities in the grazing systems and a better capacity to predict the trajectories of change following perturbations, whether physical or human. Another problem is the tendency to reach for the ‘overgrazing explanation’ when the real explanation may only be revealed by careful analysis of the case history (Rowntree et al. 2004).

Land Degradation: Prospects for Improvements So far in this chapter it has been shown that land degradation is a real and present problem in the Mediterranean. It has been demonstrated that it is a complicated (multifaceted) problem and a complex one (there are multi-stable states to be considered). Moreover it is a ‘wicked’ problem in the sense of Ludwig (in Pritchard and Sanderson 2002)—in other words it involves a host of traditional academic disciplines or social perspectives. This leads to the adoption of fixed and often rigid positions. Given this situation, it has been difficult to evaluate the two prevailing shibboleths, grazing and forest clearance, so they have grown in strength and persist up to the present day.

577

In the light of these ongoing debates, the expected crisis from global warming in the coming fifty years, and the sustained pressure on rural development, what can be done? What are the options for addressing land degradation in the Mediterranean? Evidently there are no quick and easy fixes, but future actions must take into consideration the changing conceptual and technical environments.

Technical Advances The major change in information brought about by remote sensing and computational capacity has greatly enhanced our ability to monitor the quality of the environment and track its changes through time. The CORINE Programme, for example, illustrated the huge potential of land use analysis. Perhaps greatest of all is the ability to assess the quantity and quality of vegetation and soils and especially the possibility of estimating the status of animal forage. Coupling this information with Eagleson’s theory of ecological optimality should provide an alternative strategy for estimating the annual water budget and its transformations across the Mediterranean. Also, with the increased availability of accurate and high-resolution digital elevation models, erosion modelling has moved into a new and more promising phase when coupled with vegetation cover data. Given the immense effort required to produce erosion estimates from physically based computer simulation models, they have been restricted to relatively small areas. New data and more recent models (such as the TOPMODEL, WEPP, and MEDALUS models) seem to offer fresh possibilities for obtaining estimates of erosion. They will, however, continue to need good-quality field data for calibration and validation.

Conceptual Advances Conceptually, administrators have to face up to the new paradigm of ecological complexity and its implications for understanding land degradation as outlined above. These conceptual changes have been extended, synthesized, and integrated into a metaphor (or model) for adaptive cycles of change in ecological and economic systems called ‘panarchy’ (Holling et al. 2002). The essential proposition is that there is an evolutionary cycle, comprising four phases (Figure 20.1). Starting from a collapse of the system caused by internal dynamics or external forcing, there is a phase of exploration and development as the system redevelops and reorganizes itself (the r-phase). As the system’s resources increase through time, the internal structure

578

John Thornes

of the system becomes progressively stronger, coupled, and conservative, leading to a conservative phase (the K-phase) with an optimization of available resources. These are the traditional r- and K-phases of logistic growth, dominated by the r- and K-type species of earlier ecologists. Holling et al. (2002b) have added two further phases. The release (or Ÿ) phase in which the system collapses with a large loss of resources before wholesale reorganization of the system occurs, and the restructuring (·) phase that restarts the cycle. The parallels with the boom/bust cycle modelled by Thornes and Brandt (1994) and illustrated in Figure 20.9 are to be expected from the latter’s use of the Lotka–Volterra competitive-logistic modelling strategy (Lotka 1925; Volterra 1926). In this study it was argued that the unstable cycle results from the Lotka– Volterra approach. The r- and K-parameters were tuned by Mediterranean vegetation and climate values and the model has further been used to test the stability of the grassland-tree interaction in South Africa and the emergence of hot spots in homogeneous vegetation covers (Thornes 2005). Later in their treatise, Holling et al. (2002b) show how the basic metaphors can be extended to develop cascaded cycles at different time and space scales, with leakage of resources and constraining behaviour between different levels (Figure 20.10). This new and rich paradigm introduces the full range of complexity characteristics to evolving ecosystems and nicely removes the difficulties of the conventional

log time (years)

1 cm

1m

(Clementsian) model of succession. It has important implications for ecological evolution in a land degradation context, even though it is neither fully worked out, nor is the panarchy theory fully tested. The proposed cycle is intrinsic and gives rise to periodicity through the internal dynamics and is constrained by changes in the external forcing agents that may themselves be periodic or simply stochastic. We can hypothesize that the degradation we see in the Mediterranean today is not the same as that observed by the Greek philosophers, but merely a different, later adaptive cycle. From the management point of view, the paradigm prompts the question of where we are in the cycle. The problem is to identify the critical thresholds of stability and instability, separating the zone of trajectories that lead either to complete vegetation cover or to absolutely bare soil and catastrophic erosion (Thornes 1985). Once identified, managers can decide if the system is on the brink of change when tiny adjustments might produce massive effects. When the panarchy paradigm is coupled to the new spatial ecology through the hypercycle (Dieckemann et al. 2000), real advances are in prospect. In the new spatial ecology, two- and three-dimensional spatial structures emerge and result in apparent movement of vegetation and erosional patches across the landscape. The other big shift, conceptually, is from the soil conservation approach, stressing technical fixes and practical measures, to the soil husbandry movement that stresses an integrated effort coupled to traditional

100 m 10 km 1,000 km

4

1,000 years

3

100 years

2

10 years

1

1 year

0

0.1 year

-2

0

2 4 6 log space (metres)

Fig. 20.10. A nested set of adaptive cycles in time and space (modified from Holling et al. 2002). See text for discussion.

Land Degradation

methods and public involvement. The Australian Land Care experience (Roberts 1992) that arose from the National Soil Conservation Programme catalysed rural action and provided a framework for community land care action. In 1989 it was recognized that ‘[t]here is a need for a major national integrated program, which incorporates a local, co-operative, self-help approach and the development of individual property plans and increased assistance to state and local governments within an overall framework of policy guidance. The active co-operation of landholders is pivotal to the success of the program’ (Roberts 1992: 19). This is clearly different from the heavily bureaucratic approach enshrined in the Water Framework Directive for catchment management in Europe (Chapter 8). The new technical approach to conservation lays emphasis on engineering structures that reduce surface runoff rather than attempting to reduce soil erodibility. The main shift has been towards the increased use of vegetation in soil conservation. This is now known as bioengineering. An excellent example is the study by Quinton et al. (2002) of the use of bioengineering principles for mitigating desertification. This work examines the effect of above- and below-ground plant properties on infiltration, runoff generation, and soil erosion. It then uses this information and the available literature to suggest a number of species that might prove useful in a regeneration programme. They include an extensive list of plant species, their habitats, and ecological/engineering properties. Another example of a fresh approach to conventional wisdom of the mitigation of degraded areas is the study by Margaris and Koutsidou (2002). The study, from the Greek island of Chios, is on the interaction between fire and grazing. The result shows that, on the Aegean Islands, it is clear that the productive capacity of the pastures is not enough to support the current number of animals. The proposed solution is not simply a call to reduce a mythical carrying capacity, but rather the elimination of the practice of deliberately starting fires. On the basis of their deep understanding of the system, they seek in this way to alter the form of stock farming from intensive to extensive, by subsidizing stock farmers to replace the grazed biomass by imported animal feeds. The results of a cost study show that the economic benefits of these activities could be more effectively deployed to reduce the fire hazards. Land degradation is ultimately a problem of sustainability and stability. UNEP’s Agenda 2000 required nations to move towards sustainable land use. Although Mediterranean countries, north and south, are moving

579

slowly in this direction, there is now a wide recognition of the general stability concepts that underline the search for solutions. We are entering a new, more progressive era in the land degradation problem, equipped with new paradigms and new mind-sets that bode well for the future.

References Barrocou, G., Scabbia, M. G., and Paniconi, C. (1994), Threedimensional model of salt water intrusion in the Capoterra coastal aquifer system (Sardinia). Proceedings of 13th Salt Water Intrusion Meeting. Department of Territorial Engineering, University of Calgiari, Italy. Behnke Jr., R. H., Scoones, I., and Kervan, C. (1993), Range Ecology at Disequilbrium: New Models of Natural Variability and Pastoral Adaptation in African Savannas. Overseas Development Institute, London. Blondel, J. (2006), The ‘design’ of Mediterranean landscapes: a millennial story of humans and ecological systems during the historic period. Human Ecology 34: 713–29. Brandt, J. C. (1989), The size distribution of through-fall drops under vegetation canopies. Catena 16: 507–24. Briske, D. D. and Noy-Meir, I. (1998), Plant responses to grazing: a comparative evaluation of annual and perennial grasses, in A. Ghazi and V. P. Papanastasis (eds.), The Ecological Basis of Livestock Grazing in Mediterranean Ecosystems. Office for Official Publications of the European Community, Luxembourg, 13–27. Burke, S. (1998), Groundwater over-exploitation: a case study in Castilla-La Mancha, Spain, in P. Mairota, J. B. Thornes, and N. Geeson (eds.), Atlas of Mediterranean Environments in Europe: The Desertification Context. John Wiley & Sons, Chichester, 100–1. and Thornes, J. B. (1998), Actions Taken by National Governmental and Non-Governmental Organisations to Mitigate Desertification in the Mediterranean. Concerted Action Report.1, EUR18490EN. Office for Official Publications of the European Community, Luxembourg. Conacher, A. J. and Sala, M. (1998), Land Degradation in the Mediterranean Environments of the World. John Wiley & Sons, Chichester. Cronhjort, M. B. (2002), The interplay between reaction and diffusion, in U. Dieckmann, R. Law, and J. A. J. Metz (eds.), The Geometry of Ecological Interactions. Simplifying Spatial Complexity. Cambridge Studies in Adaptive Dynamics. Cambridge University Press, Cambridge, 151–69. Despois, J. (1961), Developement de l’utilisation des terres de l’Afrique sepentrionale (avec reference a l’Espagne), in L. D. Stamp (ed.), Histoire de l’utilisation des terres de régions arides. Recherche sur la zone aride XVII. UNESCO, Paris, 245– 62. Dieckmann, U., Law. R., and Metz, J. A. J. (eds.) (2002), The Geometry of Ecological Interactions. Simplifying Spatial Complexity, Cambridge Studies in Adaptive Dynamics. Cambridge University Press, Cambridge. Elwell, H. A. and Stocking. M. A. (1976), Vegetal cover to estimate soil erosion hazard in Rhodesia. Geoderma 15: 61– 70.

580

John Thornes

Fantechi, R. and Margaris, N. S. (1986), Desertification in Europe. Reidel, Dordrecht. Francis, C. F. and Thornes, J. B (1990), Runoff hydrographs from three Mediterranean vegetation cover types, in J. B. Thornes (ed.), Vegetation and Erosion. John Wiley & Sons, Chichester, 363–85. Freschi, A. (1998), Desertification in Italy: activities at national and regional level, in Burke and Thornes (1998: 77–110). Garcia Ruiz, J. M., Lasanta, T., Ruiz-Flano, P., Marti, C., Ortigosa, L. M., and Gonzalez, C. (1994), Soil erosion and desertification as a consequence of farmland abandonment in mountain areas. UNEP Desertification Control Bulletin 25: 27–33. Gisladottir, G. (2004), Ecological disturbance and soil erosion on grazing land in south-west Iceland. in A. Conacher (ed.), Land Degradation. Kluwer, Dordrecht, 109–26. Godron, M., Guillerm, J. L., Poissonet, P., Thiault, M., and Trabaud, L. (1981), Dynamics and management of vegetation, in F. de Castri, D. W. Goodall, and R. L. Specht (eds.), Mediterranean-Type Shrublands. Elsevier, Amsterdam, 317–42. Gomes, J. L., Arrue, J. L., Lopez, M. V., Sterk, G., Richard, D., Gracia, R., Sabre, M., Gaudichet, A., and Frangi, J. P. (2003), Wind erosion in a semi-arid agricultural area of Spain, Catena 52: 235–56. Goodess, C. and Palutikof, J. (2002), Local and regional responses to global climate change in south-east Spain, in N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes. John Wiley, Chichester, 247–69. Grove, A. T. and Rackham, O. (2001), The Nature of Mediterranean Europe: An Ecological History. Yale University Press, New Haven. Gunderson, L. H. and Holling, C. S. (eds.) (2002), Panarchy: Understanding Transformations in Human and Natural Systems. Island, Washington. Holling, C. S. and Gunderson, L. H. (2002), Resilience and adaptive cycles, in Gunderson and Holling (2002: 25–63). and Ludwig, D. (2002a), In quest of a theory of adaptive change, in Gunderson and Holling (2002: 3–24). and Peterson, G. D. (2002b), Sustainability and panarchies in Gunderson and Holling (2002: 63–103). Holmgren, N. and Scheffer, M. (2001), El Niño as a window of opportunity for the restoration of degraded arid ecosystems. Ecosystems 4: 151–9. IPCC (Inter-Governmental Panel on Climate Change) (2001), Third Assessment Report. Kakembo, V. (2003), Factors affecting the invasion of Pteronia (blue bush) onto hillslopes in Ngqushwa (formerly Peddie) District, Eastern Cape. Ph.D. thesis, Rhodes University, South Africa. Kirkby, M. J. (1980), The problem, in M. J. Kirkby and R. P. C. Morgan (eds.), Soil Erosion. John Wiley & Sons, Chichester, 1–16. Köchy, M., Mathaj, M., Jeltsch, F., and Malkinson, D. (2008), Resilience of stocking capacity to changing climate in arid to Mediterranean landscapes. Regional Environmental Change 8: 73–87. Kosmas, C., Danalatos, N. G., Lopez-Bermudez, F., and RomeroDiaz, M. A. (2002), The effect of land use on soil erosion and land degradation under Mediterranean conditions, in N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 57–81.

Langbein, W. B. and Schumm, S. A. (1958), Yield of sediment in relation to mean annual precipitation. American Geophysical Union Transactions 39: 1076–84. Lavee, H., Poesen, J., and Yair, A. (1997), Evidence of high efficiency water harvesting by ancient farmers in the Negev desert. Journal of Arid Environments 35: 341–8. Le Houérou, N. (1981), The impact of man and his animals on Mediterranean vegetation, in F. De Castri, D. W. Goodall, and R. F. Specht (eds.), Mediterranean Type Shrublands. Elsevier, Amsterdam, 479–521. Lesschen, J. P., Kok, K., Verburg, P. H., and Cammeraat, L. H. (2007), Identification of vulnerable areas for gully erosion under different scenarios of land abandonment in Southeast Spain. Catena 71: 110–21. Llamas, M. R. (1988), Conflicts between wetlands conservation and groundwater exploitation: two case histories in Spain. Environmental Geology and Water Sciences 11: 241–51. Lotka, A. J. (1925), Elements of Physical Biology. Williams & Wilkins, Baltimore. Margaris, N. S. and Koutsidou, E. (2002), Landscape protection from grazing and fire, in N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 83–92. Märker, M., Angeli, L., Bottai, L., Costantini, R., Ferrari, R., Innocenti, L., and Siciliano, G. (2008), Assessment of land degradation susceptibility by scenario analysis: A case study in Southern Tuscany, Italy. Geomorphology 93: 120–9. Mtimet, A. (1999), Atlas des Sols Tunisiens. Ministry of Agriculture, Republic of Tunisia. Mulligan, M. (1998), Modelling the complexity of land surface response to climatic variability in Mediterranean environments, in M. G. Anderson and S. M. Brooks (eds.), Advances in Hillslope Processes. John Wiley & Sons, Chichester, ii. 1099–150. Nafaa, R. and Watfeh, A. (2002), Holocene and actual degradation of the environment in the Mamora Forest (Morocco). International Journal of Anthropology 15: 263–70. Naveh, Z. and Dan, J. (1973), The human degradation of Mediterranean landscapes in Israel, in F. de Castri and H. A. Mooney (eds.), Mediterranean Type Ecosystems, Springer, New York, 373–90. Obando, J. A. (2002), Modelling the impact of land abandonment on regeneration of semi-natural vegetation: A case study from the Guadalentin, in N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 419–29. Papanastasis, V. P. (1998), Livestock grazing in Mediterranean ecosystem: an historical and policy perspective, in V. P. Papanastasis and D. Peter (eds.), Ecological Basis of Livestock Grazing in Mediterranean Ecosystems, EUR18308EN. Office for Official Publications of the European Communities, Luxembourg, 5–10. (2004), Traditional vs. contemporary management of Mediterranean vegetation: the case of the island of Crete. Journal of Biological Research 1: 39–46. Price, P. N. (1998), Actions taken by Moroccan national agencies to combat desertification, in Burke and Thornes (1998: 113– 49). Pritchard Jr., L. and Sanderson, S. E. (2002), The dynamics of political discourse in seeking sustainability, in Gunderson and Holling (2002: 147–32).

Land Degradation Pueyo, Y. and Alados, C. L. (2007), Effects of fragmentation, abiotic factors and land use on vegetation recovery in a semiarid Mediterranean area. Basic and Applied Ecology 8: 158– 70. Quinton, J., Morgan, R. P. C., Archer. N. A., Hall, G. M., and Green, A. (2002), Bioengineering principles and desertification mitigation, in N. A. Geeson, J. B. Thornes, and C. J. Brandt (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses. John Wiley & Sons, Chichester, 93–105. Rackham, O. and Moody, J. A. (1996), The Making of the Cretan Landscape. Manchester, Manchester University Press. Reynolds, J. F. and Stafford Smith, D. M. (2002), Global desertification: do humans cause deserts? in J. F. Reynolds and D. M. Stafford Smith (eds.), Global Desertification: Do Humans Cause Deserts? Dahlem University Press, Berlin, 1–21. Riksen, M., Brouwer, F., and de Graaff, J. (2003), Soil conservation policy measures to control wind erosion in north-western Europe. Catena 52: 309–27. Roberts, B. (1992) Land Care Manual. University of New South Wales Press, Sydney. Rodda, J. C. (1976), Basin studies, in J. C. Rodda (ed.), Facets of Hydrology. John Wiley & Sons, Chichester, 257–98. Rowntree, K., Duma, M., Kakembo, V., and Thornes, J. B. (2004), Debunking the myth of overgrazing and soil erosion. Land Degradation and Development 15: 203–14. Salvucci, G. D. and Eagleson, P. (1992), A Test of Ecological Optimality for Semi-arid Vegetation. Report 335, Ralph M. Parsons Laboratory, MIT, Cambridge Mass. Scoones, I. (1989), Economic Development and Zimbabwe’s Communal Areas. Pastoral Development Network Paper 27b, Overseas Development Institute, London. Thirgood, J. V. (1981), Man and the Mediterranean Forest. Academic Press, New York. (1988), Goat grazing in Cyprus. Unasylva 40: 59. Thomas, D. S. G. and Middleton, N. J. (1994), Desertification: Exploding the Myth. John Wiley & Sons, Chichester. Thornes, J. B. and Brandt, C. J. (1994) Erosion-vegetation competition in a stochastic environment undergoing climatic change, in A. C. Millington and K. Pye (eds.), Environmental Change

581

in Drylands: Biogeographical and Geomorphological Perspectives. John Wiley & Sons, Chichester, 305–20. (1985), The ecology of erosion. Geography 70: 222–36. (2002), Emerging Mosaics, In N. A. Geeson, C. J. Brandt, and J. B. Thornes (eds.), Mediterranean Desertification: A Mosaic of Processes and Responses, John Wiley & Sons, Chichester, 419–29. (2005), Coupling erosion, vegetation and grazing. Land Degradation and Development 16: 127–38. and Rowntree, K. (2006), Integrated catchment management in semi-arid environments in the context of the European Water Framework Directive. Land Degradation and Development 17: 255–64. Tilman, D. (1982), Resource Competition and Community Structure. Princeton University Press, Princeton, NJ. Tsouris, G. (1985), The depletion of forests in the Mediterranean region—a historical review from ancient times to the present. Scientific Annals of the Department of Forestry and Natural Environment, KH: 281–300. UNEP (1992), Earth Summit ’92. The UN Conference on Environment and Development, Rio de Janeiro. Volterra, B. (1926), Fluctuations in the abundance of a species considered mathematically. Nature 118: 558–60. Wainwright, J. and Thornes, J. B. (2003), Environmental Issues in the Mediterranean: Processes and Perspectives from the Past and Present. Routledge, London. Warren, A. and Barring, L. (2002), Introduction, in A. Warren (ed.), Wind Erosion on Agricultural Land in Europe, EUR20370. Office for Official Publications European Community, Luxembourg. Woodward, J. C. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediments and Water Quality in River Catchments. John Wiley & Sons, Chichester, 365–89. Zachar, D. (1982), Soil Erosion. Developments in Soil Science 10. Elsevier, Amsterdam. Zorzou, M. (2004), Suspended sediment delivery and sediment properties in mountain catchments of western Greece. Ph.D thesis, University of Leeds.

This chapter should be cited as follows Thornes, J. B. (2009), Land degradation, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 563–581.

This page intentionally left blank

21

Water Resources Jean Margat

Introduction The geography of natural water resources in the Mediterranean basin cannot simply be reduced to the study of water inputs, water distribution, and the pattern of runoff-generating precipitation determined by climate and relief—although these are, of course, fundamental controls (Margat 1992; Benblidia et al. 1996). Any consideration of basin-wide water resources also needs to consider a range of territorially determined factors affecting water resources. These include: (1) the nature of surface and underground flows, which depends on river basin and hydrogeological characteristics; (2) the natural storage capacity of lakes and aquifers and their role in regulating flows, and any losses from these stores which reduce the resulting flows; (3) the existence of favourable conditions for water management and exploitation such as suitable sites for dam construction and the productivity of aquifers, as these factors dictate accessibility to water resources and the production costs; (4) the natural quality of the water, its vulnerability to pollution and its capacity for self-purification; (5) any constraints imposed for reasons of environmental conservation, which may effectively exclude a proportion of water reserves from the category of exploitable resources.

The Geography of Mediterranean Water Resources It is important to appreciate that each of these factors influences the assessment of water resources in a given area and each factor has its own geography (Margat 1997; Margat and Vallée 1999a). In spite of the broad similarities in climate and landscape between the different parts of the Mediterranean basin, there are

considerable variations between regions that impact upon the availability of water resources. Many of the factors affecting water resources cited above are subject to a similar degree of variation (Grenon and Batisse 1989; Chapter 8) and these are discussed in turn below.

The Uneven Distribution of Inputs Marking the transition between the temperate climate of Europe and the aridity of North Africa and the Near East, the Mediterranean climate contains wide variation, and this is reflected in a highly uneven distribution of rainfall (Benblidia et al. 1996; Margat and Vallée 1999a ; Chapter 3). For example, moving from one extreme to another, average annual rainfall ranges from more than 3,000 mm in parts of the Dinaric Alps to less than 50 mm in Libya. There is, first and foremost, a contrast between north and south, but east–west variations are also noticeable. In the west, the Mediterranean regions of Spain and France receive rather less rainfall than some coastal areas of the Maghreb. In the east, in Anatolia and the Levant, precipitation inputs decrease steeply from the coast towards the interior and in Egypt and Libya the Sahara desert extends right up to the coastal zone. These patterns are highlighted in Figure 21.1 which shows the basin-wide distribution of potential annual runoff or effective precipitation.

Blue and Green Water Precipitation inputs in the Mediterranean basin give rise to two types of water resource: (1) the surface and groundwater flows, or ‘blue water’, which make up the conventional natural renewable resources, and (2) the soil moisture, or ‘green water’, which can be utilized by agriculture (Falkenmark and Rockstrom 2004). Both types of water resource are unevenly distributed across the Mediterranean basin. The average yearly runoff that

584

Jean Margat

> 500 mm/yr 200 – 500 100 – 200 50 – 100 < 50

0

500 km

Fig. 21.1. Distribution of potential annual runoff (effective precipitation) in the Mediterranean basin.

forms the domestic resources of each country ranges from more than 1 million m3 per km2 in some mountainous areas in the north, to less than 10,000 m3 per km2 in the arid zones of the south, or, in terms of local rainfall inputs, it ranges from more than 1,000 mm per year to less than 10 mm (Figure 21.1). The annual average amount of ‘green water’ depends on the real level of evapotranspiration from cultivated land and varies from 700 mm at the most in the north to less than 100 mm in the south, or from 7,000 m3 to less than 1,000 m3 per hectare per year. This means that irrigation is necessary almost everywhere.

Drainage Basin Size The storage zones that collect and retain blue water are very fragmented and this resuslts in water resources being distributed between a large number of drainage basins. Such fragmentation demands a large network of independent measurement and management approaches. The Mediterranean basin is dominated by short and steep river channel systems than drain small catchments (Chapters 8 and 11). The topography of the region dictates that the headwaters of most rivers lie above 500 m in the upland zone and the watershed of river systems that drain to the Mediterranean Sea is close to the coast in many countries (Woodward 1995)

as shown in Figure 21.2. In fact, only twenty-one river basins that drain to the Mediterranean Sea cover an area of more than 10,000 km2 , and only three (the Nile, Rhône, and Po basins) cover more than 50,000 km2 (Figure 21.2). Some basins are also divided by national borders, meaning that their resources are shared by several countries. This is especially true of the Iberian and Balkan peninsulas and the Near East, the most notable case being that of the Nile basin, which is shared by ten countries, but for the most part lies outside the Mediterranean zone (Howell and Allan 1994; Woodward et al. 2007). Similarly, a great number of aquifer systems, generally limited in extent and closely linked to surface watercourses, are found mainly in alluvial or karstic and sometimes volcanic contexts. These are often formed in highly compartmentalized hydrogeological structures, with the sole exception of the sedimentary basins of the northern Sahara which impinge only slightly on the Mediterranean basin. The quantity of renewable water resources contained in these basins is, however, minimal.

Fluvial Runoff and Flow Regimes The small size of many Mediterranean drainage basins and the low level of rainfall inputs in less humid areas combine to reduce average stream flow volumes

Water Resources

585

Mean annual discharge > 1,000 m3 s-1 Mean annual discharge 100 – 1,000 m3 s-1 Other permanent rivers Seasonal or ephemeral streams

0

500 km

Fig. 21.2. The Mediterranean drainage network and river basins.

(Chapter 8). Excluding the Nile, only twenty-five watercourses (and all of them are found in the northern Mediterranean) have average flows greater than 100 m3 s−1 (Figure 21.2). Large rivers are the exception in the Mediterranean basin, where only the Rhône, the Po, and the Nile have average flow rates of over 1000 m3 s−1 (3.5 km3 per year). These rivers owe their high volumes to the fact that their catchments extend beyond the climatic zone of the Mediterranean into the Alps or equatorial Africa (Chapter 8).

Flow Regimes Owing to the generally low levels of reservoir storage, there is a great deal of variation in flow rates throughout the Mediterranean basin—both between seasons and between years. The low flows of summer are always far lower than the annual mean flow—commonly being less than one-tenth, or even one-hundredth of monthly winter flows and, where there is no contribution from groundwater, they can be zero (Chapter 8). Overall, for the whole of the Mediterranean basin, the low flows of regular surface resources amount to only a fifth of average total flows, and do not exceed 100 km3 per year. However, this irregularity of flow is not present

to the same extent in all regions and it becomes more accentuated the further south one goes, as increased aridity aggravates these hydrological contrasts (Margat 1998a ).

Surface Flows and Transmission Losses Mediterranean watercourses do not retain all the water they collect as they suffer considerable transmission losses due to evaporation or consumption by riparian vegetation (Servat et al. 2003). Their outlet flow, especially at the point of discharge into the sea, can be significantly lower than the sum of inputs into their basin. As a proportion of inputs, these losses—which are amplified by human consumption (especially through irrigation)—can, for example, reach 35 per cent in the Ebro basin, 18 per cent in the Medjerda basin in Tunisia, and 64 per cent over the entirety of the Nile basin prior to the construction of the Aswan High Dam (the real flow rate of the Nile into the Mediterranean has since been further reduced). It is therefore possible to underestimate potential resources if they are only calculated solely on the basis of river discharge measurements taken in the lower reaches of river basins.

586

Jean Margat

Natural Storage Capacity

Water Quality

The natural water reserves of the Mediterranean basin (both on the surface and in the groundwater store) are, in total terms, quite large. The Alpine lakes of the Po and Rhône basins alone have a capacity of 220 km3 , and one could add the reserves contained in glaciers to this figure (Chapter 12). However, these reserves have only a moderate regulatory effect on river flows, as the levels, and therefore the water reserves, of lakes and alluvial aquifers vary little, and karstic aquifers can empty quickly (Chapter 10). Some of the very deep, multi-level limestone karsts around the Mediterranean basin have very large storage capacities. These systems can cope with large interannual recharge variations and they often support exploitation under high pumping rates (Bakalowicz et al. 2008). The Saharan aquifers on the southern and eastern border of the Mediterranean basin (Egypt, Libya, Tunisia) that were mentioned earlier store large quantities of water, but this is essentially ‘fossil’ water with a very low rate of renewal and therefore does not contribute significantly to the regulation of river flows. No permanent watercourses exist in these arid regions and the groundwaters constitute a nonrenewable resource (Margat 1998b).

The natural quality of surface water and groundwater varies across the Mediterranean and often fluctuates over time both in respect of chemical composition and biotic content. In the north, quality is usually satisfactory for most uses, most notably for purification into drinking water, and for most aquatic ecosystems. Nitrate pollution of groundwater has been reported in many areas and is largely due to the use of nitrogen-rich fertilizers. On the island of Majorca nitrate concentrations in groundwater of up to 700 mg l−1 have been reported (Candela et al. 2007). In the catchments in the south, water quality often falls in association with falls in water quantity, so low quality frequently goes hand-in-hand with scarcity of resources. The greater water treatment costs—especially for drinking water production—that are thus necessitated to a lesser or greater extent, further accentuate the inequality in terms of available resources between the north and the south. The most common natural water quality issues, both in the north and the south, are the hardness of groundwater in carbonate-rich areas (which are found throughout the basin) and the often high turbidity of surface waters. However, it is the often high salinity of perennial water that reduces freshwater resources in southern countries. In Tunisia, for example, 26 per cent of surface water, 90 per cent of phreatic groundwater, and 80 per cent of deep-pumped groundwater has a salinity of more than 1.5 g per litre. However, it is difficult to classify the quantities of moving water in each basin in terms of quality, since the quality of water flowing within a single watercourse tends to decline as it goes downstream. The degree to which Mediterranean water, especially groundwater, is naturally protected against pollution or other threats to its quality (in other words, its vulnerability), also varies across the region. Degradation caused by humans is more widespread and affects a larger quantity of Mediterranean water than degradation by natural processes.

Water Resources and the Impact of Human Activities The Influence of Land Use The flow regimes of Mediterranean river basins, especially the smaller ones, are very sensitive to the state of the soil and of plant cover in the upstream catchment and hence to the impact of human activities that can alter these characteristics. In general, this tends to lead to an increase in the irregularity of their flows (Chapter 8). Their much diminished base flow rates greatly reduce their capacity for self-purification and increase their vulnerability to pollution. Prat and Munné (2000) examined the controls on water quantity and quality in the Congost stream of north-east Spain. The high demand for water from agriculture, the urban environment, and industry means that channel flows are effectively maintained by effluent returns from sewage plants and little or no dilution takes place from natural stream flows. Such lack of dilution has severely degraded the ecological productivity of the channel zone. Since most of the groundwater stores of alluvial plains, particularly in coastal areas, are easily accessible and exploitable, they are also vulnerable to the risk of overexploitation. They are also often poorly protected against pollution threats as explained below.

The Uneven Distribution of Water Resources Water Resources by Country The natural renewable water resources of the Mediterranean basin—in other words the sum of surface water and groundwater flows—total about 600 km3 in an average year and data for each country are presented in Table 21.1 (Margat and Vallée 1999a ). However, according to their size and their climate, countries within the basin can possess internal resources that

TABLE 21.1. Water resources in the Mediterranean basin by country and continent Countries and territories in the Mediterranean basin

Natural renewable resources (km3 /year) in an average year Internal resources 1 Surface watera

2 Groundwaterb

3 Overlapc

4 Totald 1 + 2−3

Spain France Italy Malta Slovenia Croatia Bosnia-Herzegovina Serbia-Montenegro FYR Macedonia Albania Greece

26.6 63.5 172.5 0 4.1 10.0 14.0 15.5 5.4 25.7 55.5

10.4 32.0 43.0 0.05 3.0 ∼9 ∼2.0 ∼2.0 ∼1 6.2 10.3

∼9 31.5 33 0 2.9 ∼1 ∼2 ∼1.5 ∼1 ∼5 7.8

28 64 182.5 0.05 4.2 ∼18 14 ∼16 5.42 26.9 58

Total North

392.8

118.95

—

417.1

Turkey Cyprus Syria Lebanon Israel Palestine: West Bank Gaza

External constraints: proportion of natural resources

6 Total 4+5

of external origin: index of dependence (%)

13.65 0 0 1 14.8 16.25

28.35 72.5 191.3 0.05 4.2 32 14 16 6.42 41.7 74.25

1.2 10.6 3.9 0 0 42.7 0 0 14.9 36.8 16.8

23.8e

440.9e

5 External resourcese 0.35 8.5 8.8 0

Â

20.0 0.4 2.4 3.2 0.45

∼15 0.3 ∼2 2.5 0

66 0.8 5 4.8 0.63

3.45 0 0.96 0 0.38

69.45 0.8 5.96 4.8 1.01

6.6 0 12.3 0 38.0

0.07 0

0.5 0.05

0 0

0.57 0.05

0 0.01

0.57 0.06

0.0 16.7

77.8

2.8e

80.6e

37.4

−

Egypti Natural Res. Real Res. Libya Tunisia Algeria Morocco

0.3 0.3 0.2 2.9 11.64 4.9

0.5 8j 0.6 1.15 1.33 1

0 7.5j 0.1 ∼0.35 ∼1 ∼0.9

0.8 0.8 0.7 3.7 11.97 5

84 55.5 0 0.32 0.03 0

85 56.3 0.7 4.02 12 5

Total South

19.94

4.58

—

22.17

55.5e

77.7e

Overall

483.4

160.9

—

517.1

e

82

599.2e

0.1 0.9 0 0 92.2 0 97 82.5 100 0 2.7

21.4 40 110 0.015g 2.2 9 7 8 3 13 29

Present production capacity of non-conventional water (km3 /year) Regeneration of used water available for reuse 0.2 —

Reusable drainage water

∼0.2

0.0016

— — — — —

Desalination

0.13 — 0.058 0.022 —

97 0 97–8

— — —

— 0.013

96

242.6

61.0 0.7 4.6 4.1 0.18

Total East

flowing into neighbouring countries (restrictions on freedom of use) (%)

Estimated exploitable resourcesf (km3 /an)

0.3 0 29h 10.6 1.0 100 0

40 0.54 4 2.18 0.75 0.53 0.05

—

Â

— 0.011

98

0.034

0.02 0.27

94 96

0.007 0.085 0.03

0.027

00 99

Â

48.05 98.8 98.6 0 8.0 0.21 0

0 0 0 0 2.2 0.75

49.5 0.65 1.8 6 3.4

0.7 0.07 0.026

12.6 98 97

0.27 0.054 0.06 0

99 00 90

61.35 352

a Total runoff; b Groundwater recharge; c Overlap: groundwater flows collected or fed by watercourses. Equivalent to the difference between the value of column 2 and the groundwater that flows directly into the sea or across borders, or evaporates in depressions in arid zones; d Total without double counting; e Surface or groundwater flow from a neighbouring country, whether or not it adjoins the Mediterranean. The external resources of each country are not simply added to obtain the subregion figure because of sharing between countries. These figures are therefore free of double counting, and refer only to water brought in from non-neighbouring countries. The figures for the total resources in subregions and in the entire basin (column 6) take this into account; f According to criteria appropriate to each country. These figures are therefore not homogeneous; g Malta: water that is exploitable without upsetting the freshwater/saltwater balance; h Syria: on the basis of the natural outflow of the Orontes into Turkey; i Egypt: the distinction between natural and real resources is crucial here on account of the importance of inflows from the Nile; j Egypt: additional groundwater recharge by irrigation in the Nile valley and delta: 7.5 km3 /year of ‘secondary resources’. Source: National sources compiled by le Plan Bleu (2002).

588

Jean Margat

Fig. 21.3. Natural renewable and exploitable water resources per country in the Mediterranean basin (annual means).

amount to literally billions of cubic metres per year for the better off, or just millions of cubic metres for the most deprived. Thus, the three countries with the greatest water wealth in the Mediterranean basin, namely France, Italy, and Turkey, together possess well over half (60%) of the internal water resources of the whole basin. By contrast, the least well-off parts of the basin (Malta, Gaza, Cyprus, and Libya) have only 3 per cent of this total (Figure 21.3 and Table 21.1). To these internal resources should be added the external resources brought by cross-border rivers, as the Mediterranean basin extends to some countries that do not adjoin the Mediterranean Sea itself. While these resources have a secondary role in Western Europe and the Maghreb, they are more important in south-eastern Europe (Albania, Croatia, Greece) and are, naturally, crucial in Egypt, where the proportion of water resources that originate from abroad is overwhelming (98.8%). The predominance of external resources in the south-east of the Mediterranean is essentially due to the contribution of the Nile to Egypt as shown in Table 21.2. The differences in the exploitability of resources further deepen the divisions between countries.

Water Resources by Population Calculating the water resources available per head of present-day populations provides deeper insights into the levels of water wealth or poverty in the Mediterranean basins of each country. The distribution is again highly uneven (Figure 21.4), ranging from extreme water poverty (less than 100 m3 per year in Gaza, less than 200 m3 per year in Israel, Libya, and Malta), to abundance (more than 10,000 m3 per year in Albania, Bosnia-Herzegovina, Croatia, Slovenia, and SerbiaMontenegro). At the time of writing, 114 million Mediterranean people (45% of the total population of the basin), spread TABLE 21.2. Key figures on internal and external natural and exploitable water resources for the three main regions of the Mediterranean basin Subregion of the basin North (Europe) East (Near East) South (Africa) TOTAL

Natural internal % Real external Total resources resources resources (km3 /year) (km3 /year) (km3 /year) 417 78 22 517

81 15 4 100

24 3 55.5 82.5

441 81 77.5 599.5

%

73.5 13.5 13 100

Water Resources

589

Fig. 21.4. Natural renewable water resources and real exploitable water resources per inhabitant (in 2000) of each country in the Mediterranean basin.

over eight countries, live under the water poverty level of 1,000 m3 per year in natural water resources (annual average) per inhabitant. Below this level, tensions generally arise between requirements and resources, particularly in countries where irrigation is necessary. In six countries or territories (with a total of 48 million inhabitants) resources per head fall below even the level of absolute ‘penury’ of 500 m3 per year: Israel, Libya, Malta, the Palestinian territories, Tunisia, and Algeria. Furthermore, several of these countries are either already using almost the entirety of the exploitable resources in their sector of the Mediterranean basin, or have exceeded this level, thereby necessitating the use of imports from outside the basin (Israel, Libya). This inequality becomes even more acute when differences in levels of socio-economic development are considered, as it is often in the poorest countries that resources are scarcest and therefore most expensive to exploit (see Oki and Kanae 2006).

Evaluating and Exploiting Water Resources Irregular inflows and outflows, limited monitoring networks for both surface water and groundwater,

reductions in the streamflow of many watercourses, exchanges between interdependent reserves of surface water and groundwater, together with the growing impact of human activities on water regimes, make it very difficult to obtain the necessary hydrological data that would form the basis of an evaluation of the resources of natural and renewable water in the Mediterranean basin. Gathering these data requires that a larger number of observations be made over longer periods than is the case in other regions, and also that appropriate corrections be applied to stream discharge measurements. As perennial surface water flows that may be exploited by direct withdrawal generally only constitute a small proportion of surface water resources, it is often necessary to implement regulatory schemes in order to manage these resources. The distribution of dams across the different areas of the Mediterranean basin is uneven, and the sites of these dams vary with regards to their profitability, ease of maintenance, and their location in relation to points of water use. The potential for controlling water supply throughout the year by using very large capacity reservoirs is very limited. Few of the reservoirs so far constructed in the Mediterranean have a capacity approaching 1 km3 , and there is only

590

Jean Margat

one giant reservoir in the basin: the Aswan High Dam reservoir (Lake Nasser) on the Nile (164 km3 ), but this is an unusual example and is way beyond the southern limit of the Mediterranean region. In the whole of the Mediterranean basin, the total quantity of water held in artificial reservoirs (not counting Lake Nasser) currently stands at just 66 km3 . However, the potential capacity of reservoirs that could still be created tends not to be included in these calculations, as the feasibility criteria of these projects are complex and subject to change. There are many potential sites for small reservoirs in various northern and southern (especially in Morocco, Tunisia, and Algeria) countries, but their usefulness as regulatory sources and their useful lifespan are both reduced by the risk of sedimentation (Lahlou 1988). Sources of groundwater vary in size, and their accessibility and exploitability depend upon the depth and productivity of the aquifers in question. Furthermore, they often, particularly in the north, feed perennial surface water flows. The complications that arise from this interdependence place serious limits on their exploitability, and may also generate conflicts between users. In several countries, a proportion of groundwater flows directly into the sea via coastal or submarine karst springs that are difficult to tap; over the Mediterranean as a whole, these flows account for 17 per cent of the basin’s groundwater (Chapter 10; Bakalowicz et al. 2003; Fleury et al. 2008). The often considerable difficulties involved in water management limit the proportion of truly usable natural resources. The irregularity of an often overwhelming proportion of the water flows in the region necessitates the construction of water management systems that vary greatly in cost and efficiency. In addition to this, the decline in efficiency that accompanies the expansion of these systems and of water exploitation schemes causes costs to rise. The efficiency of artificial reservoirs is reduced by the high evaporation rate to which their contents are subject (from 0.8 m per year to more than 2 m per year). In Algeria, for example, evaporation from reservoirs stands at between 1.3 and 2.2 m per year. In Egypt, evaporation causes the loss, on average, of 10 billion m3 per year from Lake Nasser and this is equivalent to about 12 per cent of the controlled flow of the Nile. The position of water management installations in relation to points of use—such as towns and irrigation schemes—often leads to high transport and pumping costs. Using criteria that are as much economic as social in nature, it can be argued that not all Mediterranean water can be counted as a resource because people are either unable or unwilling to exploit part of that resource. A proportion of non-perennial surface water,

most notably in the case of heavy flooding, cannot be controlled, either because of the lack of a suitable site for a dam, or because to control it would necessitate the construction of installations that would be impractical on grounds of cost or environmental impact. A proportion of perennial surface water and groundwater is also unusable because it is needed for on-site applications or for the preservation of wetland ecosystems (Chapter 9).

Competition for Water Resources Water, like any resource, is the object of rivalries and even of conflicts between the growing needs of users, all the more so where it is in short supply. Competition between abstractors of groundwater and surface water users frequently arises, often being driven by ignorance of the interdependence between these two components of a single resource. Groundwater abstraction can cause springs to dry up and reduce the flow rates of streams in dry periods, while dams can prevent the recharge of aquifers. The largest area of competition is between users in towns and their rural counterparts. The provision of drinking water in cities, generally a priority in social and political terms, is set against irrigation, the area of use normally hardest hit by demand management measures due to its predominant position among the various sectors of current water consumption. All the users and river basin managers also need to consider the ecological requirements of the aquatic system itself (Prat and Munné 2000). The Mediterranean environment was previously richer in perennial streams and wetlands than it is today (Chapter 9). Over time human activities have caused these to recede, whether unintentionally, as a consequence of water withdrawal, or voluntarily, through the drainage of marshes judged to be unhealthy or the reclamation of cultivable land. In Spain, in recent years, for example, 40,000 hectares of wetlands have been lost—three-quarters of which were linked to intensively exploited groundwater resources. Contemporary ecological concerns, in particular with regard to the preservation of the aquatic environment and wetland zones, which are becoming increasingly rare, have necessitated the designation of ‘restricted outflows’, which can impose local restrictions, varying according to national environmental policies, on water management and withdrawal. In this way a proportion of surface and groundwater is excluded from the total quantity of exploitable resources. The flashy discharge peaks of Mediterranean river systems mean that flooding is a major problem, and

Water Resources

this hazard is magnified by the increasing concentration of population and human activities in high-risk zones (Chapter 18). The risks are often underestimated due to complacency or the short-term interests of those occupying the land in question. However, in the north and in the south, a string of recent disasters, which are not as ‘natural’ as some might argue, serve as a reminder that water in the Mediterranean is not only a resource but also a potential danger. In many Mediterranean basins the prevention of water-related risks is one of the principal objectives of water management.

Threats to Mediterranean Water Resources Environmental changes caused by the people of the Mediterranean are reducing the renewal rate and the quality of water resources (Bethemont 2000; Jeftic et al. 1996). Water renewal cycles are threatened by various types of land use that have been transforming the Mediterranean environment for centuries. This process of transformation has accelerated in the twentieth century. Rampant urbanization (particularly in coastal areas) and deforestation in the south, along with the cessation of slope maintenance through terraced agriculture in the north, can all lead to enhanced runoff volumes and sediment transfer (Chapter 8). In combination with the engineering of stream channels, these changes to river basins have made stream flows less predictable and, it can be argued, have reduced the total quantity of reliable water resources. Groundwater levels are very sensitive to the effects of over-exploitation, especially when abstraction is carried out by disparate groups who do not coordinate their activities, and whose objectives are short-term in nature. This is a particular problem in coastal areas, where the equilibrium between groundwater and seawater can easily be disrupted, leading to virtually irreversible salt water contamination of coastal aquifers. This has happened in most Mediterranean countries, but the effects can be strongly conditioned by local geological and hydrological factors, as Calvache and PulidoBosch (1997) have shown for three coastal aquifers in southern Spain. The quality of surface waters and groundwaters can be degraded by various sources of pollution. A range of pollution types can be identified, including point sources of effluent returns, industrial waste, and accidents, and diffuse sources from the use of fertilizers and pesticides in intensive agriculture, or poor waste management. Groundwater is more vulnerable in this respect as,

591

while cleaning up groundwater pollution is more timeconsuming, any loss of groundwater quality may impact upon perennial surface watercourses. Groundwater-fed springs are commonly the best source of potable drinking water in much of the Mediterranean and the bottled spring water trade is an important industry. Surface water stored in reservoirs, as well as in natural lakes, is also under threat from eutrophication triggered by land use and climatic factors (e.g. Reed et al. 2008; Chapter 9). Pollution can wipe out a proportion of water resources, rendering them unusable, or push up drinking water purification costs to a prohibitive level. Pollution is not the sole preserve of industrialized northern countries, in which it is undoubtedly more prevalent, but also dealt with more effectively. It is becoming more common in the south, where underdeveloped purification systems and preventive measures increase the likelihood of pollution events, and the consequences of these events are aggravated by the scarcity and frequently low natural quality of perennial water resources.

The Sustainability of Water Management Measures The high suspended sediment loads found in floodwaters in the Mediterranean basin, especially in the south (Chapter 8), causes the silting-up of reservoirs to be a particular water resource problem. It shortens the useful lifespan of the reservoirs as regards flow regulation, in spite of the enormous ‘sleeping reserves’ that they were designed to hold. In the south, the loss rate of useful reservoir capacity currently runs at between 0.5 and 1 per cent per year, sometimes higher: 0.5–2 per cent in Algeria, where the lifespan of average-capacity reservoirs is between thirty and fifty years; 0.5 per cent in Morocco, where the reduction of regulatory capacity attributable to silting is currently equivalent to a loss of irrigation potential of between 6,000 and 8,000 hectares per year; and, finally, 1–2.5 per cent in Tunisia. Algeria’s reservoirs have already lost 25 per cent of their total initial capacity, while Morocco’s had lost 8 per cent (800 million m3 ) of theirs by 1990, with certain reservoirs already being half-filled by silt (Lahlou 1988; Chapter 8). Since only a limited number of sites are suitable for the construction of dam and reservoir installations, and some of these have already been developed, the development and silting up of all of these reservoir sites is to be expected in the longer term, probably before the end of the twenty-first century. Efforts at prevention through the stabilization of hillslopes and the installation of

592

Jean Margat

sediment traps can at best stall this process, but cannot prolong reservoir life indefinitely in such high sediment yield environments. A significant decline in the amount of water resources available from these forms of regulatory management is thus inevitable. As regards the exploitation of the non-renewable resources that are provided by a number of large aquifer reservoirs in several southern countries (Libya, Tunisia, Algeria), especially in the Saharan regions on the outer rim of the Mediterranean basin, the necessarily limited duration of exploitation is governed, like any drilling operation, by the rate of extraction that has been selected, and will at best be in the order of fifty years. In addition to this, the quality of the water extracted from these aquifers can be degraded through contamination by saline water even before the reserves are exhausted, further shortening their exploitable lifespan.

The Threat of Climate Change The water resources of the Mediterranean basin are not protected from the effects of any climate change caused by the intensification of the greenhouse effect over the twenty-first century (Margat 2004). In the south, the risk of the climate becoming more arid is inescapable, and this would have the dual effect of reducing water resources and increasing water requirements as a result of higher evaporation rates and lower rainfall. In the north, the danger is rather that the climate will become more extreme, wetter in the winter and drier in the summer, and more unpredictable (Chapter 3). This would also have consequences for patterns of water use, possibly raising water requirements in the summer, while increasing the risk of flooding. Water resources are not simply a raw material provided by nature to be exploited until it can no longer renew itself. They are vulnerable to the impact of the various effects of their exploitation and of water use. They are also vulnerable to the effects of the various lifestyles and activities of the human population. Mediterranean water resources are naturally fragile, but this fragility is increased by intensive exploitation and by the activities of Mediterranean people, which have radically altered the natural water budget. The greater the use made of these resources, the more vulnerable this use becomes to unpredictable natural factors such as droughts and falls in water quality, and to direct human-induced pressures. The water resources of today are not, it would seem, quite what they were in the past even if the standards of evaluation have themselves evolved. The evaluation

of water resources in the Mediterranean basin needs in particular to combine the hydrological data necessary to predict their future potential with feasibility studies of their sustainable exploitation. All this should be coupled with an understanding of their sensitivity to environmental factors that takes into account the limits imposed by the need for their conservation. Two opposing tendencies could lead to the revision of calculations of exploitable water resources in the Mediterranean basin. On the one hand resources may increase as, in the face of diminishing availability, the pressure of demand forces up the ceiling of the water management and exploitation costs that are judged to be acceptable (within the limits, nonetheless, of countries’ differing economic capacities and the physical limits of natural resources). Alternatively, resources may drop, owing to an increasing awareness of the limits imposed on exploitation by the need for environmental conservation. The final outcome may vary according to the countries in question, differing water and environmental policies, and changes in the overall situation.

A Diminishing Resource At present, pressures on resources vary, but are often intense. In the Mediterranean basin, the geography of water use and hence of water requirements is no less varied than that of water resources, although there are certain dominant trends such as the predominance of irrigation and the increasing requirements of cities and of the tourist industry. Throughout history, demand for water has been met exclusively by the management and exploitation of natural water resources. As demand has increased, so these management measures and withdrawals have intensified. The levels of channel diversions, control of irregular flows by dams and reservoirs, and exploitation of groundwater by tapping or pumping vary, but in several countries these are already high. The growing scarcity of sites with development potential, the reduction in the flow of perennial streams, and also the intensive exploitation of groundwater, with all its undesirable consequences, are key problems for water resource planners in the region. The scarcer resources become the more sensitive they are both to quantity-related pressures resulting from withdrawals and quality-related pressures caused by the recirculation of used water. Although the different aspects of these pressures throughout the Mediterranean basin result in a somewhat varied picture, the pressure on resources has already reached high levels in most areas. In the Mediterranean basin as a whole,

Water Resources TABLE 21.3. Annual water withdrawal volumes for the three main regions of the Mediterranean basin Subregion in the basin

North (Europe) East (Asia Minor, Near-East) South (Africa)

Total draw-off (km3 /year)

Proportion of natural renewable resources (average exploitation index) (%)

90 18 69

19 22 88

177 billion m3 of water are withdrawn and used annually and this amounts to a quarter of the average level of natural resources, and certainly a larger proportion of exploitable resources. The total volume of annual water withdrawals is shown in Table 21.3 for the European, African, and Asian sectors of the Mediterranean basin. However, withdrawal levels and pressures on resources differ considerably according to subregions and even more so according to countries and river basins (Figure 21.5). In several countries, the quantity of withdrawals from the basin is approaching or indeed has equalled the quantity of resources in an average year, and therefore

593

threatens to exceed it in a dry year. The exploitation indexes for renewable water calculated per country in the basin still show considerable variation, but nonetheless exceed 50 per cent in eight countries (Table 21.4). They therefore indicate that tense situations exist, at least at the local level and according to local circumstances. In some countries the maximum level has been reached, approaching or sometimes even exceeding 100 per cent. This may indicate a loss of equilibrium or the deliberate but unsustainable recourse to non-renewable resources (Libya), or the fact that a proportion of the resources is being used more than once, such as in the collection and reuse of used water (Israel), or the recirculation of drainage water (Egypt). Final natural resource consumption figures, which represent the quantity of water that is withdrawn and then not replaced after being used, are also high in Mediterranean countries, where two factors help to raise them: (1) the relative importance of high-consumption agricultural uses and (2) the high proportion of used water from towns, industries, and tourist developments that, as a consequence of coastal population concentration, is discharged into the sea and therefore reduces the amount of water returned to streams or groundwater.

75 – 100% 50 – 75% 25 – 50% 0– 25%

0

500 km

Fig. 21.5. The exploitation index of natural renewable water resources across the Mediterranean basin. This is the ratio of withdrawals to the sum of internal and external resources (see Table 21.4).

594

Jean Margat TABLE 21.4. Present-day pressures on water resources in the Mediterranean basin

Countries and territories in the basin

Spain France Italy Malta Slovenia Croatia Bosnia-Herzegovina Serbia-Montenegro FYR Macedonia Albania Greece

Date Total

Surface water

Groundwater

1997 1999 1998 1997–8 1996 1996 1995 1995 1996 1995 1997

17.94 16.67 42 0.025 0.03 0.19 0.1 0.8 1.85 1.40 8.7

14.71 14.72 31.6

10.25 5.0 18 0.004f

0.02 0.04 0.02 0.5 1.65 0.77 5.14

3.23 1.95 10.4 0.025 0.01 0.15 0.08 0.3 0.2 0.63 3.56

89.7

69.17

20.54

41.2

1997 2000 1997 1996 1999 1994 1996

11.1 0.295 3.85 1.3 1.12 0.13 0.13 17.93

10.25

7.68

10.1

1995–6 1998 1996 2000 1998

59.5 2.0 2.27 2.9 1.9

55g 0.1 1.15 2.0 1.7

4.5 1.9 1.12 0.9 0.2

35.1 1.3 1.1 1.7 0.9

Total North Turkey Cyprus Syria Lebanon Israel Gaza West Bank Total East Egypt Libya Tunisia Algeria Morocco Total South OVERALL

Final consumption (km3 /year)a

Withdrawals (km3 /year)

68.57 176.2

Â

6.3 0.15 2.85 0.9 0.05

 Â

59.95 139.4

4.8 0.145 1.0 0.40 1.07 0.13 0.13

Â

0.15 0.1 0.3 0.8 0.6 6.0 Mean %

6.0 0.2 1.6 1.0 1.1 0.1 0.07

Indicators of pressure on water resources (%) Exploitation indexb

Final consumption indexc

84 42 38.2 162d 1.5 2 1.4 10 60 10 29

48 13 12.5 67f 0.15 0.8 0.6 3.4 25 4.5 17

18.7 28 55 96e 80 120 260 25

Mean %

21.7 96 233 69 41 56

8.6 15 41 40 47 110 108 9.4 12.2 110 200 42 23 34

8.62

40.1

Mean %

87.9

51.4

36.83

91.4

Mean %

27.5

14.3

Notes: a Net consumption by use plus used water not returned to continental waters (discharged into the sea); b Exploitation index: annual withdrawals/average yearly flow of total natural renewable resources in %; c Consumption index: annual final consumption/average yearly flow of total natural renewable resources in %; d Relates to resources that can be exploited without breaking the freshwater/saltwater balance; e Syria: relates to real resources; f Malta: includes returns of used water of non-conventional origin (desalination); g Egypt: gross withdrawals, including recovery of drainage water. Sources: National and international figures compiled by le Plan Bleu (2002).

At present, final resource consumption over the whole of the Mediterranean basin is probably near to 100 billion m3 per year, or 70 per cent of the total draw-off, 55 per cent of which occurs in the southern (40.1 km3 per year) and eastern (10.1 km3 per year) countries (Table 21.4). Naturally, these quantity-related pressures are accompanied by the impact of discharges of used water upon water quality as well as that of other sources of pollution. In an average year, around 20 billion m3 , a significant proportion of which is not purified, are released into the continental waters of the Mediterranean basin. While they are less industrialized than northern countries, southern countries have scarcer water resources, and so may suffer the effects of

this pollution to a greater degree. In Egypt, for example, where industries discharge 550 m3 of waste water per year (57% of all by-products) into the Nile, the number of deaths attributable to illnesses caused by water pollution in the Nile have, according to the WHO, risen from 19,395 in 1979 to 48,458 in 1987. The efforts undertaken to improve sanitation, the purification of used water, waste management infrastructure, and the general prevention of pollution have not, by and large, been as extensive as those put into expanding water supplies. This is also the case in other parts of the world, but the consequences in the Mediterranean basin are more serious because its water resources are scarcer, and have greater demands placed upon them.

Water Resources

Temporal Trends in Water Resource Pressures The rate of increase of these pressures on water resources in the Mediterranean has varied, and continues to vary. The available national statistics do not allow a reliable historical account of water withdrawal to be established, not even for the period following the middle of the twentieth century. Notwithstanding the variation in their quality, the figures for the last twenty-five years revealed by these statistics are significant (Margat and Vallée 1999a). In Mediterranean countries as a whole, withdrawals have risen overall by 50 per cent, although this varies according to region, with increases of nearly 30 per cent in the north, where growth has slowed, of more than 60 per cent in the south, and a doubling of withdrawals in the east. The differences between individual countries are also considerable and will involve: (1) a tripling of withdrawals in Libya; (2) a near-doubling in Algeria, Spain, France, Syria, and Turkey; (3) increases of more than 50 per cent in Greece, Lebanon, Morocco, and Tunisia; (4) moderate growth in Egypt, and stabilization in Italy, where withdrawals are very high; (5) low growth or stabilization in Israel and Cyprus, where the ceiling of resources has been reached. In the Mediterranean basin, specific withdrawal histories are difficult to establish, but the general tendencies would seem to be largely convergent and exploitation figures have typically risen in the same proportions in many areas (Margat and Vallée 1999a, b). The pressures on resources will continue to grow in the twenty-first century, although to a lesser extent than most national and international projections have suggested (e.g. Vision Mediterraneenne de 2000). While future demand might be moderated by measures to improve the economy of water use, it will still, with a few exceptions, essentially be covered by the exploitation of natural resources. According to the studies carried out as part of the Plan Bleu project (2003), water demand and the withdrawals that it necessitates could grow in total across the Mediterranean basin, following current trends, by about 20 km3 per year between 1995 and 2025. Such an increase would give an overall total of 211.9 km3 per year by 2025 (Table 21.5). This amounts to a rise of just 11 per cent with considerable differences between subregions: a 4 per cent drop in the north, and rises of 20 per cent in the south and 45 per cent in the east. The highest growth is predicted to occur in the Mediterranean parts of Libya (+65%), Turkey (+60%), and Algeria (+50%). These trends would further widen the disparity between north and south.

595

TABLE 21.5. Water demand predictions for 2025 in the three regions of the Mediterranean basin Subregions in the Mediterranean basin North East South TOTAL

Trend projections of total water demand in 2025 (km3 /year) 86.2 27.8 97.9 211.9

Generally, then, it is in the countries where pressure on resources is already high that it is likely to increase the most. According to these projections of current trends, the exploitation indexes for natural renewable resources will be above the 50 per cent mark in the Mediterranean basins of nine countries in 2025, and will be approaching or will have already exceeded 100 per cent in Egypt, Israel, and Gaza. When applied to exploitable resources, these indexes would rise to 50 per cent in thirteen countries, while approaching or exceeding 100 per cent in eight countries. Water penury threatens the twenty-first century. Even according to the relatively optimistic hypothesis that natural water resources are sustainable (with the effects of climate change not being felt before the middle of the twentyfirst century); structural water penury would still be expected within a generation across a substantial part of the Mediterranean basin as a result of increased pressure on resources. Two further projections that show rising levels of water penury support this conclusion: first, there will be a drop in the level of water resources per inhabitant as a result of projected demographic expansion. Per-inhabitant water resources (average renewable resources) of below 1,000 m3 and 500 m3 are, respectively, accepted indicators of water poverty and absolute water penury. Second, given the decline in per capita resources predicted to occur between now and 2025, we are left with the prospect of 164 million Mediterranean people in eight countries living in a situation of water poverty (compared to 114 million at present). Six of these countries, with a combined population of 67 million people, will be in a situation of water penury (compared to 48 million at present). By 2025 more than half the population of the Mediterranean basin will therefore be in a situation of stress or penury in relation to natural water resources, let alone in relation to the resources that can actually be utilized. There will be a significant reduction in the net availability of water (resources remaining after withdrawals and recirculated water have been taken into account) in terms both of quantity and quality. This availability will

596

Jean Margat

decline at a more or less rapid rate as a function of final consumption, and will tend to disappear completely in those countries and places where it is already very low (Malta, Israel, Gaza, and Libya). In this way, we can begin to sketch out the geography of present and future water penury in the Mediterranean basin. This reveals the same kinds of disparities as those found in studies of water resources and the pressures upon them. It can be said with certainty that the countries or territories that are under the greatest threat are: Malta, Gaza, Israel, and Libya, where structural penury is already present and then Cyprus, Tunisia, Syria, Algeria, and Egypt, where occasional instances of penury will become more frequent, and regional structural penury will become the norm.

Conclusions In common with other parts of the world (see Pimentel et al. 1997), naturally occurring water resources are distributed unevenly in the Mediterranean basin, while the ease with which they can be exploited to meet demand also varies markedly. These are major contributing factors to the disparity of development possibilities and cost levels in the provision of water supplies in the Mediterranean basin. These costs will, in general, increase, although once again unevenly and rising most steeply where they are already high. This will further handicap the development of the southern and eastern countries—especially in irrigated agriculture— more than that of northern countries. This will distort competition in the Euro-Mediterranean free-trade zone scheduled to begin in 2010. In an analysis of water scarcity and food import in six Mediterranean countries (Algeria, Egypt, Israel, Libya, Morocco, and Tunisia), Yang and Zehnder (2002) argue that water scarcity is a rigid limiting factor to food production and that food imports are imperative for compensating water resource deficiency. This situation increases the potential for conflicts of use, whether between sectors of use (principally between urban communities and agricultural irrigation), between the different regions of a country (upstream–downstream rivalry, regional opposition to water-transfer projects), or between countries forced to share resources such as in the Iberian Peninsula, the Balkans, the Near East, and the Nile basin (Maury 1990). Levels of water resources and water availability vary between neighbouring territories or countries, and this opens up the possibility of creating transfer schemes in order to alleviate situations of water

penury, whether between regions within a single country (these are already planned or in operation in Spain, France, Greece, Israel, Egypt, Libya, Tunisia, Cyprus, and Morocco), or between different countries, creating an international water trade. Indeed, projects between France and Spain (Catalonia), Albania and Italy, Turkey and the Near East have been proposed. The geography of water resources represents a key control on development and biodiversity in the Mediterranean basin and the pressures to which they are now subjected are expected to increase during the course of the twenty-first century. This poses major challenges for river basin planners across the region and for the implementation of the Water Framework Directive in the European Mediterranean. Given the population and climate change predictions for the region, the geography of present and future situations of water penury will become increasingly important in the economic geography and geopolitics of the Mediterranean world.

References Bakalowicz, M., Fleury, P., Jouvencel, B., Promé, J. J., Becker, P., Carlin, T., Dörfliger, N., Seidel, J. L., and Sergent, P. (2003), Coastal karst aquifers in Mediterranean regions: a methodology for exploring, exploiting and monitoring sub-marine springs. Tecnologia de la Intrusion de Agua de Mar en Acuiferos Costeros: Paises Mediterraneos. IGME, Madrid. El Hakim, M., and El-Hajj, A. (2008), Karst groundwater resources in the countries of eastern Mediterranean: the example of Lebanon. Environmental Geology 54: 597–604. Benblidia, M., Margat, J., and Vallée, D. (1996), L’Eau en région méditerranéenne—Water in the Mediterranean Region. Conférence Euroméditerranéenne sur la gestion de l’eau, Marseilles, November 1996; new edn. 1997, Blue Plan, Sophia Antipolis. Bethemont, J. (2000), La question de l’eau en Méditerranée. Revue de l’économie Méridionale, 48/191: 179–90. Calvache, M. L. and Pulido-Bosch, A. (1997), Effects of geology and human activity on the dynamics of salt-water intrusion in three coastal aquifers in southern Spain. Environmental Geology 30: 215–23. Candela, L., Wallis, K. J., and Mateos, R. M. (2007), Nonpoint pollution of groundwater from agricultural activities in Mediterranean Spain: the Balearic Islands case study. Environmental Geology 54: 587–95. Falkenmark, M. and Rockstrom, J. (2004), Balancing Water for Humans and Nature: A New Approach in Ecohydrology. Earthscan, London. Fleury, P., Bakalowicz, M., de Marsily, G., and Cortes, J. M. (2008), Functioning of a coastal karstic system with a submarine outlet, in southern Spain. Hydrogeology Journal 16: 75–85. Grenon, M. and Batisse, M. (eds.) (1989), Futures for the Mediterranean Basin: The Blue Plan. Oxford, Oxford University Press.

Water Resources Howell, P. P. and Allan, J. A. (eds.) (1994), The Nile: Sharing a Scarce Resource. Cambridge University Press, Cambridge. Jeftic, L., Keckes, S., and Pernetta, J. (1996), Climate Change and the Mediterranean. Edward Arnold, London, ii. Lahlou, A. (1988), The silting of Moroccan dams, in M. P. Bordas and D. E. Walling (eds.), Sediment Budgets. IAHS Publication 174: 71–7. Margat, J. (1992), L’Eau dans le bassin méditerranéen. Situation et prospective, Blue Plan 6. Éditions Economica, Paris. (1997), Cadre Géographique des Ressources en Eau dans la Région Méditerranéenne. Rapport au congrès international: L’acqua nei paesi mediterranei – Problemi di gestione di una risorsa scarsa. Consiglio Naz. delle Ricerche/Naples, 4–5 December 1997. CNR/IREM. Il Mulino, Bologna, 31–52. (1998a ), Sécheresses et ressources en eau en Méditerranée. Rapport à la conférence sur la politique de l’eau en Méditerranée, Valencia Espagne 16–18 August, 1998: Session Gestion des sécheresses. CR édités par le Réseau Méditerranéen de l’Eau, Madrid. (1998b), Les Eaux souterraines dans le bassin méditerranéen. Resources et utilisations. Documents du BRGM 282. Blue Plan– BRGM, Orléans. (2004) L’Eau des Méditerranéens. Situation et perspectives. Rapport Techniques du PAM, à paraître. and Vallée, D. (1999a ), The Mediterranean in Figures. Water Resources and Uses in the Mediterranean Countries. Figures and Facts. Blue Plan, Sophia Antipolis. (1999b), Vision méditerranéenne sur l’eau, la population et l’environnement au XXe siècle Mediterranean Vision for Water, Population and the Environment in the 21st Century. MEDTAC, document pour le IIe Forum mondial de l’eau de la Haye. Global Water Partnership, Conseil Mondial de l’Eau. Blue Plan, Sophia Antipolis.

597

Maury, R.G. (1990), L’Eau dans les pays méditerranéens de l’Europe communautaire. Études méditerranéennes 15. Centre Interuniversitaire d’Études Méditerranéennes, Poitiers. Oki, T. and Kanae, S. (2006), Global hydrological cycles and world water resources. Science 313: 1068–72. Pimentel, D., Houser, J., Preiss, E., White, O., Fang, H., Mesnick, L., Barsky, T., Tariche, S., Schreck, J., and Alpert, S. (1997), Water resources: agriculture, the environment, and society. BioScience 47: 97–106. Prat, N. and Munné, A. (2000), Water use and quality and stream flow in a Mediterranean stream. Water Research 34: 3876–81. Reed, J. M., Leng, M. L., Ryan, S., Black, S., Altinsaçli, S., and Griffiths, H. I. (2008), Recent habitat degradation in karstic Lake Uluabat, western Turkey: A coupled limnological– palaeolimnological approach. Biological Conservation 141: 2765–83. Servat, E., Najem, W., Leduc, C., and Shakeel, A. (2003), Hydrology of Mediterranean and Semi-arid Regions. IAHS Publication 278. International Association of Hydrological Sciences, Wallingford. Woodward, J. C. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water Quality in River Catchments. John Wiley & Sons, Chichester, 365–89. Macklin, M. G., Krom, M. D., and Williams, M.A. J. (2007), The Nile: Evolution, Quaternary river environments and material fluxes, in A. Gupta (ed.), Large Rivers: Geomorphology and Management. John Wiley & Sons, Chichester, 261–92. Yang, H. and Zehnder, A. J. B. (2002), Water scarcity and food import: a case study for southern Mediterranean countries. World Development 30: 1413–30.

This chapter should be cited as follows Margat, J. (2009) Water resources, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 583–597.

This page intentionally left blank

22

Air Pollution and Climate Jos Lelieveld

Introduction It has long been known that atmospheric pollutants can be hazardous to human health and ecosystems. This includes effects from episodic peak levels as well as the long-term exposure to relatively moderate concentration enhancements. Environmental issues related to air pollution include acidification, mostly by the strong acids from sulphur and nitrogen oxides, eutrophication by the deposition of reactive nitrogen compounds, the reduction of air quality by photo-oxidants and particulate matter, and the radiative forcing of climate by increasing greenhouse gases and by aerosol particles. Many air pollutants are photochemically formed within the atmosphere from emissions by traffic, energy generation, industry, the burning of wastes, and forest fires. The Mediterranean basin in summer is largely cloudfree, and the relatively intense solar radiation promotes the photochemical formation of ozone (O3 ) and peroxyacetyl nitrate (PAN); O3 being health hazardous at levels in excess of about 100 Ïg/m3 . Ozone is formed in the lower atmosphere as a by-product in the oxidation of reactive carbon compounds such as carbon monoxide (CO) and non-methane volatile organic compounds (NMVOC), catalysed by nitrogen oxides (NOx ≡ NO + NO2 ). In summer, notably the period from June to August, transport pathways of air pollution near the earth’s surface are typically dominated by northerly winds, carrying photo-oxidants and aerosol particles from Europe into the Mediterranean basin. Aerosol particles with a diameter of less than ∼10 Ïm (PM10 ) can have adverse health effects at a concentration of about 30 Ïg/m3 or higher. The fine mode particles (<2 Ïm diameter) are mostly composed of sulphates, nitrates, and particulate organic matter, whereas the

coarse mode particles (≥2 Ïm) often contain substantial amounts of sea salt, Saharan dust (Chapter 14), and other mineral components. The aerosols can form widespread hazes that scatter and absorb solar radiation, thus reducing downward energy transfer and surface heating. Increased aerosol scattering causes a negative radiative forcing of climate (cooling tendency), to be weighted against the positive radiative forcing (warming tendency) by increasing greenhouse gases such as carbon dioxide (CO2 ), methane (CH4 ), nitrous oxide (N2 O), halocarbons, and tropospheric ozone (IPCC 2001). Note that the radiative forcing by aerosol and ozone is mostly regional in nature because the lifetime of these compounds is relatively short (days to weeks), whereas long-lived greenhouse gases CO2 and CH4 (years to centuries) are mixed more or less homogenously on a global scale. This has the important implication that air quality control measures, e.g. reducing sulphate particles, have near-instantaneous effects on climate forcing, whereas control measures aimed at CO2 have a much more delayed response. Since the regional cooling tendency by aerosol scattering can mask the enhanced greenhouse effect by CO2 , air pollution emission controls can have a complex and possibly unexpected outcome. In Europe the abatement of air pollution has been promoted by the UN Convention on Long-Range Transport of Air Pollution (LRTAP), which was signed in 1979 and came into force in 1983. Since 1980 the emissions of air pollutants in Europe have indeed decreased, as illustrated in Table 22.1. Substantial emission reductions have been achieved, for example nearly 50 per cent for CO and more than 75 per cent for SO2 . Nevertheless, the emissions have decreased relatively less in the Mediterranean member states of the European Union (EU). In Greece, Portugal, and Spain, for example, the

600

Jos Lelieveld

TABLE 22.1. Air pollution emissions in Europe in the year 2000, and emission reductions between 1980 and 2000

SO2 50.0

CO

NOx (as NO2 )

SO2

11.1 −32%

11.5 −30%

8.7 −76%

France, Greece, Italy, Portugal, and Spain Emissions (Tg/year) 16.8 5.3 Emission change 1980–2000 −41% −19%

4.7 −9%

3.6 −65%

0.5

Source: Vestreng et al. 2004; expert data from , accessed 14 November 2008.

0.1

EU 25 Emissions (Tg/year) 36.6 Emission change 1980–2000 −48%

NMVOC

10.0 5.0 1.0

anthropogenic emissions of NMVOC and NOx have actually increased and those of sulphur dioxide (SO2 ) have hardly decreased during this period. Table 22.1 also shows that five Mediterranean EU countries contribute nearly half to the EU25 air pollution emissions. The spatial distribution of emissions for the year 2000 in Europe are presented in Figure 22.1. Furthermore, in southern Europe biogenic emissions of NMVOC, e.g. isoprene and monoterpenes, can substantially enhance the formation of photochemical air pollution (Barros et al. 2003). This chapter concentrates to some degree on the results of the Mediterranean Intensive Oxidant Study (MINOS), based on a measurement campaign using two aircraft and a ground station in the summer of 2001, as presented in a special issue of the journal of Atmospheric Chemistry and Physics edited by N. Mihalopoulis and M. de Reus (available from <www. atmos-chem-phys.org>). The study domain, with a focus on the eastern Mediterranean region, as observed from the NASA SeaWiFS satellite in August 2001, is shown in Figure 22.2. The campaign was performed using Crete as an operating base to study the long-range transport and photochemistry of air pollution, including the effects on air quality and climate. Some of the main results have been presented by Lelieveld et al. (2002a). The foremost conclusion from the MINOS studies is that summertime air pollution not only adversely affects human health throughout the Mediterranean basin, but also that aerosol particles strongly influence the radiation energy budget in the area, perhaps conspiring with some of the regional climate consequences of global atmospheric change.

Climate and Meteorology The Mediterranean climate is characterized by humid winters with cyclonic storms, and warm, dry summers, with occasional extended drought periods (Bolle 2003; Chapter 3). The average north–south temperature

NOx 50.0 10.0 5.0 1.0 0.5 0.1

NMVOC 50.0 10.0 5.0 1.0 0.5 0.1

Fig. 22.1. European air pollution emissions, 2000, in thousand tonnes per 50 km grid cell (after Lövblad et al. 2004). NOx emissions in units of NO2 .

gradient across the basin, from the Alps to North Africa, is remarkably large, about 25◦ C. In some locations, e.g. along the Adriatic coast, precipitation is among the highest in Europe (∼1,000 mm per year) and in the mountains of the Balkan Peninsula it is typically >2,000 mm per year, whereas in much of North

Air Pollution and Climate

601

Fig. 22.2. Widespread aerosol haze in the Mediterranean basin, as observed from the SeaWiFS satellite during the MINOS campaign in August 2001 (© Orbimage, NASA).

Africa this can be more than an order of magnitude less. The Mediterranean weather is strongly influenced by the geographical positions of the Azores high and the Icelandic low, which can modify the mean westerly flow (Traub et al. 2003). When the high and the low are strongly developed, and therefore the meridional pressure gradient is strong, relatively moist air masses are transported to Europe. In the alternate case, the zonal flow is weak, and blocking weather systems prevail over central Europe. The interannually varying pressure gradient between the Icelandic low and the Azores high pressure systems is known as the North Atlantic Oscillation (NAO), an indicator of the inten-

sity of synoptic weather systems over the North Atlantic Ocean. Even though the NAO is a natural mode of climate variability, it can be influenced by anthropogenic climate change (Corti et al. 1999). It is linked to the general atmospheric and oceanic circulation systems, the latter being affected by the formation of cold bottom water in the Arctic Ocean and the influx of salty water from the Mediterranean Sea through the Strait of Gibraltar (Bolle 2003) (Chapter 2). By definition, a relatively strong pressure gradient between the Azores high and the Icelandic low yields a positive NAO index, a relatively weak gradient a negative NAO index. Although its influence has been studied primarily for the winter season,

602

Jos Lelieveld

the NAO is associated with considerable monthly and interannual variability, and effects have been identified for all seasons. During the MINOS campaign in August 2001, the mean NAO index was 0.3. In summer, the InterTropical Convergence Zone (ITCZ) and the Azores high shift to higher latitudes. The Azores anticyclone, in combination with eastward moving low pressure systems over central Europe, lead to westerly flow in the lower and middle troposphere towards the Mediterranean basin. Spreading of the high pressure across central Europe in summer weakens the westerly flow, and European air is mostly transported to the basin by northerly winds, at right angles with the strong east–west pressure gradient near the surface. In the upper troposphere, air masses are transported in the westerlies from the North Atlantic Ocean (Figure 22.3). During spring and summer the Asian continent heats up, developing a heat low over northern India, Pakistan, and Iran. This causes the ITCZ to move north, generating the Indian monsoon with intense deep convection. In the upper troposphere, a semi-permanent anticyclone is located over the surface heat low, called the Tibetan high. The tropical easterly jet stream is an inherent feature of the Indian summer monsoon and the Tibetan high. It is a belt of strong easterly winds at the southern periphery of the upper tropospheric anticyclone. The tropical easterly jet, between 100 and 200 hPa pressure altitude, transports air from Asia over northern Africa and, aided by the upper tropospheric anticyclone over the Arabian Peninsula, towards the eastern Mediterranean (Traub et al. 2003). Under the influence of the subtropical jet stream and the westerlies, the air is subsequently transported back to the east over the Asian continent. During summer the Mediterranean basin is directly under the descending branch of the Hadley circulation, driven by deep convection in the ITCZ. In the upper troposphere, poleward moving air from the ITCZ is deflected by the Coriolis force. The resulting westerly flow reaches a maximum in the subtropical jet stream at about 40◦ N in summer, north of the Mediterranean Sea. As a result of subsidence, the region is characterized by dry and cloud-free conditions with high solar radiation intensity. Over land, convection can develop, which is generally not very deep at coastal locations. The air near the surface travels to the equatorial region as the synoptic northerly flow combines with the trade winds, carrying moisture evaporated from the Mediterranean Sea southward. In summer, teleconnections between the Mediterranean region, the Indian monsoon, and the Sahel rainfall regimes can be important, varying on an interannual

Upper troposphere

Middle troposphere

Boundary layer

500 km Fig. 22.3. Schematic of air flows during the MINOS campaign in August 2001. The arrows indicate atmospheric transport during approximately 2–4 days (after Lelieveld et al. 2002a).

time scale (Ward 1992). The atmospheric pressure at sea level in the eastern Mediterranean is anticorrelated with that in the Indian monsoon, mainly in the July– September period, while that in the western Mediterranean is positively correlated, with a maximum during September–November. The meridional wind component over the central and eastern Mediterranean basin is anticorrelated with that in the Indian monsoon. This means that a more active monsoon is connected with lower atmospheric pressure at sea level over the eastern

Air Pollution and Climate

part of the basin and higher pressure over the western basin, and stronger northerly winds over most of the Mediterranean. The northerly flow in the basin near the surface during summer, in combination with the lack of precipitation, promotes the long-range transport of pollutants from Europe (Luria et al, 1996). Air pollutants released in southern Europe may even reach the ITCZ over Africa on a time scale of 4–6 days (Kallos et al. 1998). Since the Mediterranean basin is largely surrounded by mountains, the local meteorology can be strongly influenced by orographical flows with strong diurnal cycles (Milláan et al. 1997). These flows can combine with land–sea breeze systems, so that relatively deep land-inward surface winds can be established covering distances up to 60–100 km. Since many pollutant sources are located in coastal regions, these circulations can carry air pollution into convective-orographical ‘chimneys’ that connect the surface flow to the return flow above, i.e. linking into the regional northerlies (Millán et al. 2005). Similarly, daytime orographical flows in the Alps can vent air pollution from the valleys into the lower free troposphere, where these air masses can be transported towards the Mediterranean (Henne et al. 2004). Further downwind over the sea, the signatures of these local circulation systems are evident as distinct pollution layers in the lower ∼4 km of the atmosphere, maintained over long distances owing to the stable stratification in the region. The following section addresses the long-distance air pollution transport that can be distinguished in the Mediterranean upper troposphere. This is followed by a discussion of the atmospheric chemistry and transport pathways near the surface, including the consequences for air quality and climate. The final section discusses the links between air quality, climate, and the water cycle in view of atmospheric changes expected in the twenty-first century.

Upper Tropospheric Pollution Plume In summer, the Tibetan high in the upper troposphere combines with the high over the Arabian Peninsula into a large anticyclone, which carries air from the Indian monsoon region over Africa to the eastern Mediterranean. The thunderstorm clouds in the ITCZ region over southern Asia can carry both moisture and air pollution from the surface to high altitudes on a timescale

603

of several hours, thus injecting the pollutants into the upper tropospheric anticyclone. During MINOS the impacts of this feature have been detected over Crete and the Aegean Sea, for example, by enhanced concentrations of CO, CH4 , and non-methane hydrocarbons in an extensive plume extending in altitude from the middle troposphere (6–8 km) up to the tropopause (∼15 km). Tracer-transport modelling studies indicate that the upper tropospheric pollution plume from South Asia was present for about 25 per cent of the time, whereas alternatively it was deflected to the south during which the Mediterranean upper troposphere is influenced more strongly by air mass transports from North America and the North Atlantic region (Lawrence et al. 2003). Air mass back trajectory calculations combined with global tracer transport modelling furthermore shows that the transport time of the South Asian plume, for example, from the polluted megacities in India to the eastern Mediterranean upper troposphere, is about 10–15 days (Figure 22.4; Traub et al. 2003). Within the South Asian plume, CH4 concentrations are relatively high, probably as a result of wetland and agricultural emissions, mostly from rice fields. Carbon monoxide and non-methane hydrocarbons are also enhanced (Scheeren et al. 2003). The relatively high concentrations of PAN and the products from biomass combustion (such as C2 H2 , CH3 CN, and CH3 Cl), point to the importance of biofuel use, agricultural waste burning, and other incomplete burning processes in the South Asian source region. By combining trajectory modelling and chemical analyses, east–west gradients of air pollutants in the upper troposphere over the Mediterranean basin become apparent, associated with air pollution sources in South Asia and North America, respectively. For example, the biomass burning products were significantly higher towards the east, whereas fossil fuel combustion products (including CO2 ) were significantly higher towards the west. Also HCFC-134a, a halocarbon used in air-conditioning systems in automobiles in North America, was clearly enhanced towards the western part of the basin. The chemical ‘age’ of air pollutants in the South Asian upper tropospheric plume has been estimated from the enhancement or emission ratio of biomass burning species (X) relative to CO, X/CO, which changes as a function of the travel time and the photochemical lifetime of the species involved. The calculated emission ratios have been compared with those measured downwind of India in the Indian Ocean Experiment (INDOEX) in 1999 (Scheeren et al. 2002), providing additional evidence of the South Asian origin of the air pollution over

604

Jos Lelieveld

Days 60°N

16 14 12

40°N 10 8 20°N

6 4 2

0° 0 20°W

20°E

60°E

100°E

Fig. 22.4. Transport-time spectrum showing the period between release of air pollutants at the surface in northern India and arrival in the upper troposphere (200 hPa) in August 2001 (M. Traub, personal communication).

the Mediterranean in the upper troposphere as well as the plume travel time of 10–15 days. The importance of the Asian summer monsoon in pollution transport—note that this is the region with the strongest growing emissions in the world—is not limited to the upper troposphere, but can even extend into the lower stratosphere. Over the Mediterranean basin in summer the tropopause strongly slopes downwards, from about 16 km over North Africa to about 11 km over southern Europe. It appears that in summer the westerly wind in the polar front jet and the southerly wind at the eastern flank of the upper tropospheric anticyclone converge near the tropopause (Traub and Lelieveld 2003). The consequent acceleration of the flow increases the vertical and horizontal wind shear, creating a jet streak within the subtropical jet stream and deep tropopause folds. The strong wind shear leads to clearair-turbulence and mixing of the pollutants across the tropopause.

The residence time of the pollutants in the lower stratosphere is about 2 weeks (5–30 days) after which these substances are transported back into the troposphere. Transport is in an easterly direction from the Mediterranean over the Black Sea and Asia towards the Pacific Ocean. Hence the combination of the Asian monsoon convection and the unique meteorology in the upper troposphere causes South Asian emissions to reach the lower stratosphere in summer, and contribute to the efficient long-distance transport of air pollution.

Ozone Transport and In Situ Formation Although the upper tropospheric pollution plume from South Asia is not very rich in ozone, the 4–6-km thick layer in the middle troposphere beneath this

Air Pollution and Climate

plume contains surprisingly high O3 mixing ratios up to about 240 Ïg/m3 (typically 130–200 Ïg/m3 ). By using a chemistry-general circulation model, Roelofs et al. (2003) have analysed the photochemistry and transport phenomena involved. It was found that the upper tropospheric anticyclone influences the potential temperature and vorticity distributions over the Mediterranean basin such that air is efficiently transported downward and south-eastward. Thus a large fraction (∼80%) of the air mass trajectories in the westerly wind regime, ending over the Mediterranean between 4 and 8 km altitude, originate in the upper troposphere over the North Atlantic Ocean (Traub et al. 2003). Interestingly, these trajectories are strongly influenced by stratosphere–troposphere exchange associated with cyclonic activity over the North Atlantic Ocean. As a consequence, nearly 30 per cent of the tropospheric ozone over the Mediterranean basin in summer originates from the stratosphere. In winter this fraction may even be larger (Kentarchos et al. 2001). In spite of the importance of stratosphere–troposphere exchange for the middle troposphere, near the surface the ozone is almost entirely produced by in situ photochemical production. Owing to the stable stratification of the Mediterranean troposphere, exchange between the marine boundary layer and the free troposphere is suppressed. The model calculations suggest that about 90 per cent of the O3 near the surface is locally produced, of which about 80 per cent is caused by the photochemical conversion of anthropogenic emissions transported in the northerly flow from Europe. The remainder is from natural ozone sources. Note that a substantial part of the NMVOC acting as an O3 precursor is of natural origin. Some Mediterranean tree species appear to be particularly efficient emitters of highly reactive monoterpenes (Kesselmeier and Staudt 1999; Simon et al. 2005). The combination of these reactive NMVOC and NOx from traffic and industrial sources can give rise to a rapid build-up of ozone. Not surprisingly considering the northerly flow, the western Mediterranean is mostly influenced by emissions in Western Europe, while the eastern Mediterranean is most strongly influenced by emission in Eastern Europe. The depth of the boundary layer over the Mediterranean Sea in summer is about 0.8–1 km, over which O3 mixing ratios exceed 100 Ïg/m3 during most of the summer season, and the diurnal variation is only small, about 10 per cent. These relatively high mixing ratios in background air seem to be typical for the entire

605

basin, as shown by measurements (Millán et al. 2000; Gangioti et al. 2001; Kouvarakis et al. 2000, 2002a; Nolle et al. 2002) as well as model results, which also indicate that Mediterranean ozone concentrations are among the highest in the Northern Hemisphere (Figure 22.5). Above the boundary layer the stable reservoir layer channels air pollution southwards. This pollution also originates in Europe, lofted to 1–4 km altitude by land–sea breeze circulations, shallow convection, and orographic flows (Jiménez et al. 2006). It may be expected that this layer is broken up by low-level convection and turbulence further downwind over Africa, mixing the pollution towards the surface, or transporting it further into the ITCZ (Kallos et al. 1998). The enhanced O3 levels recorded over the Mediterranean basin during MINOS have also been detected from space by the Global Ozone Monitoring Experiment (GOME) on the ERS-2 satellite. It appears that these remote sensing observations compare favourably with in situ measurements by aircraft and model results (Ladstätter-Weißenmayer et al. 2003). The coincident GOME and aircraft measurements of the O3 precursor gases NO2 and HCHO furthermore confirm the largescale importance of photochemical O3 formation from pollution sources (Kormann et al. 2003). These measurements point to very active photochemistry in the Mediterranean basin associated with strong oxidant production. It thus appears that O3 levels in the basin exceed the European Union eight-hourly air quality limit of 110 Ïg/m3 throughout most of the summer—and this is caused by European air pollution. In several parts of the basin, for example in the north-east of Spain, the EU plant protection threshold (80 Ïg/m) is exceeded more that 80 per cent of the time (Ribas and Peñuelas 2004). It is very difficult to control the ozone air quality locally, especially in the densely populated areas along the coast, since urban emissions add to the already high background O3 mixing ratios. In fact, based on scenario simulations for the year 2025, as shown in the right hand panel of Figure 22.5, it may be expected that tropospheric O3 will further increase, and that the Mediterranean region will remain as one of the ozone ‘hot spots’ in the Northern Hemisphere (Lelieveld and Dentener 2000). In these scenario calculations for 2025 it has been assumed that European NOx emissions will remain fairly constant, based on the expectation that technological improvements (such as limits on traffic NOx emissions), are compensated by an increasing

606

Jos Lelieveld May–August (present)

60°N

0°

60°S

180°

90°W

0°

90°E

180°

May–August (2025)

60°N

0°

60°S

mgm3 O3

180°

90°W

0°

90°E

180°

mgm3 O3

10

20

30

40

50

60

80

100

120

number of vehicles and a growing fossil energy consumption. Remarkably, even though the European NOx emissions may not increase in the future, the ozone concentrations over the Mediterranean may nevertheless

140

Fig. 22.5. Model-calculated ozone at the surface, averaged over the period May to August, for the present and the possible future (2025) atmosphere (after Lelieveld and Dentener, 2000).

increase as a consequence of enhanced background levels, caused by the rapidly growing ozone formation in Asian pollution air being transported on a hemispheric scale.

Air Pollution and Climate

High Atmospheric Oxidation Capacity Ozone, being a greenhouse gas and a prime oxidant, is also the main precursor of other oxidants such as the hydroxyl radical (OH), playing a key role in the photochemical degradation of species such as CO, hydrocarbons, and various sulphur compounds. Radical reaction chains can be initiated by the photo-dissociation of O3 , followed by primary OH formation from the reaction of excited oxygen atoms with water vapour (O1 D + H2 O → 2OH). The reaction chains are propagated through the catalytic action of nitrogen oxides (NOx ), which recycles the OH radicals (through HO2 + NO → OH + NO2 ) and additionally produces O3 by the photo-dissociation of NO2 . Since the primary formation of OH is controlled by ultraviolet radiation, ozone, and water vapour (all of which are abundant over the Mediterranean Sea in summer), and because OH recycling by NOx is also relatively efficient, the OH concentrations are quite high. Berresheim et al. (2003) measured an average OH concentration of 4.5(±1.1) × 106 molecules per cm3 during the MINOS campaign on Crete, with a strong diurnal cycle and high daytime peak values, as shown in Figure 22.6. This mean OH concentration is about a factor of three higher than the typical values for this latitude in summer. The associated concentration of sulphuric acid (H2 SO4 ), formed from SO2 + OH, is also quite high—up to 9 × 107 molecules per cm3 and this is

Day

among the highest reported in the literature (Bardouki et al. 2003). Note that H2 SO4 is the prime precursor of sulphate aerosol particles, discussed in the next section. During the day oxidation processes are dominated by OH radicals. However, during the night the OH concentration is very low so that atmospheric chemical pathways depend on other oxidants such as nitrate radicals (NO3 ) formed from the reaction between NO2 and O 3 . Sunlight photodissociates the NO3 radicals within seconds, so that its concentration follows an opposite diurnal cycle from that of OH (Figure 22.6). The night-time mean NO3 concentration during MINOS was 1.1(±1.1) × 108 molecules per cm3 . This is more than an order of magnitude higher than the daytime mean OH concentration of 8.2(±1.6) × 106 molecules cm3 (Vrekoussis et al. 2004). Since the reaction rate coefficients of non-methane hydrocarbons with NO3 are between 5 and 1,000 times slower than with OH, some species are preferentially converted by NO3 radicals. Vrekousis et al. (2004) calculated that, diurnally averaged, the conversion of dimethyl sulphide (DMS) by NO3 is only 33 per cent slower than that by OH, and that the further reaction of NO3 into N2 O5 and its heterogeneous conversion into HNO3 accounts for more than 20 per cent of the total nitrate production. It is interesting to see in Figure 22.6 that the variability of OH during the day is substantially less than the variability of NO3 during the night. This is because OH concentrations are ‘buffered’ through a balance between primary formation (O1 D + H2 O) and recycling through

Nigh t

Day

2.0

1.6

1.6

1.2

1.2

0.8

0.8

0.4

0.4

0

0 12

15

18

21 00 03 Loca l time (hr) (=UT C+3) Mean NO3

06 Mean OH

09

12

OH (107 molecules/cm3)

NO3 (108 molecules/cm3)

2.0

607

Fig. 22.6. Mean diurnal cycles of OH and NO3 radicals, measured at Finokalia, Crete, during the MINOS campaign in August 2001 (after Vrekoussis et al. 2004).

608

Jos Lelieveld

the catalytic action of NOx , even though the former process is dominant in the Mediterranean boundary layer (Lelieveld et al. 2002b). In summary, the oxidation capacity of the Mediterranean atmosphere is relatively high, during both day and night. The coincidence of reactive carbon compounds and nitrogen oxides, and the favourable conditions for oxidant formation promote the photochemical transformation of primary air pollutants such as NMVOC and SO2 . Although the lifetime of primary pollutant gases is thus limited, it leads to the build-up of reaction products such as ozone and aerosol precursors, which are inefficiently removed owing to the relatively slow dry deposition over the sea and the scarcity of rain. Therefore, the Mediterranean basin in summer, with its local pollution (re)circulation systems in a stable stratified atmosphere, is a ‘cooking vessel’ of photochemical smog.

Extensive Aerosol Haze Based on satellite measurements, the Mediterranean Sea has been identified as one of the maritime regions in the world with the highest aerosol optical depths (Husar et al. 1997). The earliest atmospheric chemistry measurements on Crete indicated relatively high concentrations of sulphate and nitrate, attributed to long-distance transport of air pollution (Mihalopoulos

et al. 1997). Upwind in northern Greece high levels of SO2 have been observed, attributed to coal burning in Central and Eastern Europe (Zerefos et al. 2000; Formenti et al. 2001). A small though significant fraction of 5–25 per cent of the sulphate, however, originates from natural DMS emissions by marine phytoplankton (Ganor et al. 2000; Kouvarakis and Mihalopoulos 2002; Chapter 2). Furthermore, desert dust intrusions from Africa, and additional—though smaller—fluxes of mineral dust from the Near and Middle East substantially contribute to the aerosol column (Dayan et al. 1991; Formenti et al. 2001; Chapter 14). Over Africa, the dust is often lifted to altitudes well above the Mediterranean boundary layer in synoptic disturbances. Hence, after its northerly transport from Africa it can entrain into the westerly flow over the Mediterranean in the free troposphere. It appears that the interannual variability of aerosol optical depth over the Mediterranean is strongly affected by the atmospheric column abundance of mineral dust, which correlates with the phase of the North Atlantic Oscillation that influences the atmospheric transport regime (Moulin et al. 1997). During the MINOS campaign, the fine aerosol mass consisted of more than one-third of sulphate and nearly one-third of particulate organic matter (POM), and it included substantial fractions of ammonium, black carbon, and other compounds (Figure 22.7). The sulphate was not fully neutralized and mostly present as

Fig. 22.7. Mean particle composition during the MINOS campaign for the fine (D < 2 Ïm) and the coarse (D ≥ 2 Ïm) mode aerosol based on measurements in Finokalia, Crete. The fine aerosol mass is 16.6 Ïg per m3 and the coarse aerosol mass is 24.2 Ïg per m3 (after Lelieveld et al. 2002a). NIS represents non-identified species.

Air Pollution and Climate

ammonium bisulphate—as also observed by Kouvarakis et al. (2002b). The coarse aerosol fraction mostly consisted of mineral dust and sea salt, including significant fractions of nitrate and sulphate. Lelieveld et al. (2002a) estimated that the fine aerosol particle mass was about 80–90 per cent anthropogenic, whereas about 60–80 per cent of the coarse mode particle mass was of natural origin. The pollutant aerosol sources include urban, industrial, agricultural, and forest fire emissions. Fire maps from the MODIS satellite instrument (Moderate Resolution Imaging Spectroradiometer) show that, during MINOS, west and north of the Black Sea, in Bulgaria, Romania, and in the Ukraine and Russia, respectively, extensive biomass burning occurred, 2,000 km or more upwind of Crete. These emissions reached the measurement site at Finokalia in about four days (Sciare et al. 2003a). Aircraft and mountain-based (1500 m a.s.l.) measurements corroborated the presence of pollutants in a reservoir layer above the boundary layer, in which the aerosol particles originated from Western, Central, and Eastern Europe. In general, agricultural burning and forest fires—mostly anthropogenically ignited—in the Mediterranean region contribute strongly to the aerosol burden, especially during dry spells (Sciare et al. 2003b; Chapter 19). Black carbon can originate both from biomass and fossil fuel combustion. Sciare et al. (2003a) estimated that during MINOS more than half the black carbon originated from fossil fuel use and the remainder from biomass burning. The relatively large mass of POM, on the other hand, included many oxygenated hydrocarbons, indicative of aged photochemically processed air pollution and biomass burning aerosol (Schneider et al. 2004). Only a small part of the POM appears to be related to natural emissions and this is discernible in the particles as formate, acetate, and ≥ c9 compounds, which were present only in very small concentrations. The mixture of natural and human-made particles forms an extensive haze over the Mediterranean Sea that can clearly be discerned on satellite images (Figure 22.2). Often visibility is limited, even on distant islands. Note that Crete represents the southernmost part of Europe in the Mediterranean region, remote from pollution emissions on the continent, whereas the observed local aerosol concentration is close to the European Union air quality standard for particulate matter (PM10 ) of 55 Ïg/m3 . This concentration is not allowed to be exceeded for more than thirty-five days per year. The high concentrations in Crete may be considered indicative for the even stronger violations of the air quality standards further upwind.

Top of Atmosphere

DQ W/m2

609

Surface

GHG 3.0

Aerosol -6.6

Aerosol -17.9 Fig. 22.8. Diurnal mean, clear-sky radiative forcing (Q) during MINOS (after Markowicz et al. 2002). GHG is the greenhouse gas forcing.

By performing altitude resolved radiometric measurements, Markowicz et al. (2002) inferred that the diurnally averaged shortwave radiative forcing at the surface by aerosol particles is −17.9 ± 2.1 W/m2 (Figure 22.8). This is mainly due to solar radiation scattering and absorption by particles in the fine mode aerosol, substantially reducing the surface heating. The radiative forcing at the surface compares to a forcing of −6.6 ± 2.1 W/m2 at the top of the atmosphere (TOA); the latter representing the overall loss of solar energy to space through aerosol radiation backscattering. The difference between the TOA and the surface forcing corresponds to an atmospheric heating of 11.3 ± 3.8 W/m2 , caused by the absorption of solar radiation, in which black carbon plays a key role (Sateesh and Ramanathan 2000). Note that the above-mentioned surface forcing of nearly 18 W/m2 during MINOS is somewhat less than that derived from satellite measurements over a thirtyyear period (25 W/m2 ), with a relatively larger forcing during summer than during winter (Tragou and Lascaratos 2003). This could imply that the aerosol forcing may have decreased in the 1980s and 1990s, consistent with estimates of SO2 emissions in Europe as shown in the upper part of Figure 22.9. Figure 22.8 contrasts the negative aerosol forcing with the positive radiative forcing over the Mediterranean exerted by greenhouse gases. Obviously, these strong radiative

610

Jos Lelieveld 60

(a) 50

Mtons SO2

40 30 20 Mylona (1996)

10

Van Aardenne et al. (2001) EMEP (2002)

0

(b) DSST (°C)

0.5

0

- 0.5 1930 1940 1950 1960 1970 1980 1990 2000 Year

Fig. 22.9. (a) Estimated historical SO2 emissions in Europe; (b) 5-year running mean of Mediterranean SST anomalies (after Lelieveld et al. 2002b).

perturbations may be expected to influence the Mediterranean atmospheric heating profile and the moisture budget through changes in evaporation and cloudiness.

Air Quality, Climate, and the Water Cycle The solar radiation energy input into the Mediterranean Sea, as reduced by aerosol particles, constitutes a major driving force of the oceanic thermohaline circulation (Tragou and Lascaratos 2003; Chapter 2). Indeed, substantial changes in Mediterranean deep water formation have been observed, possibly affecting salinity transport to the North Atlantic Ocean through the Strait of Gibraltar (Lascaratos et al. 1999). It should be further investigated to what extent these changes have been caused by surface radiative forcings of aerosol pollution and regional rainfall anomalies and how they link to global climate change, for example, through the North Atlantic Oscillation (NAO). The Mediterranean basin is characterized by large climate gradients, and there is concern that climate change

will be associated with the intensification of extreme weather conditions (Chapter 18). Regional European climate scenario calculations for the twenty-first century indicate a relatively strong warming as compared to the rest of Europe (Schär et al. 2004), and the number of very hot days, in particular, may increase (de Castro et al. 2004; Chapter 3). Air quality is strongly correlated with the mean temperature through the occurrence of hot and stagnant anticyclonic conditions (Lin et al. 2001) and these are expected to increase in future (Stott et al. 2004). In addition, the hemispheric background level of air pollution may increase through, for example, long-distance transport of emissions from Asia (Lelieveld and Dentener 2000). Furthermore, since a positive phase of the NAO correlates with increased transatlantic air pollution transport, a possible future change in the NAO may have consequences for air quality both in North America and Europe (Li et al. 2002). Climate scenario calculations also indicate that precipitation may increase in the western and northern parts of the basin, more often released in torrential rain events (Chapter 18), while in the drier southern and eastern parts precipitation is expected to decrease even further (Millán et al. 2005). However, the models used did not account for the effects of aerosol particles on the energy and moisture budgets. As indicated above, the strong aerosol scattering and absorption in the Mediterranean basin reduces the surface heating and thus the sea surface temperature (SST). The lower part of Figure 22.9 shows that the long-term SST variability, as influenced by the NAO, was fairly regular in the period 1930–70. Subsequently a strong cooling phase occurred, which correlates with SO2 emissions in Europe and presumably also with the sulphate aerosol burden over the Mediterranean Sea. Since the solar energy absorbed by the sea is largely returned to the atmosphere through evaporation, the negative radiative forcing at the surface, caused by sulphates and other particulate matter, suppresses evaporation and atmospheric moisture transport. To assess the possible consequences of this aerosol effect on the regional water cycle, Lelieveld et al. (2002b) performed a sensitivity study by prescribing observed low and high SSTs as boundary conditions to a general circulation model. The results demonstrate that the Mediterranean SST is a sensitive influence on the amount of precipitation received downwind in, for example, the Middle East and the eastern Sahel zone (Figure 22.10). First, the negative SST anomalies appear to correspond to drought periods in North Africa in the 1970s (Long et al. 2000). Second, the positive SST anomalies in the 1990s correlate with a recovery from

Air Pollution and Climate

611

Precipitation change: low–high SST (%) 0

0

0

20

40°N

10 0

0 -10

0

0

-20

20°N 0

-30 -40

0

0° 20°W

0°

20°E

40°E

these spells in the same period (Nicholson et al. 2000) and these are coincident with decreasing SO2 emissions and sulphate concentrations (Figure 22.9). Further studies will be needed to substantiate these links between aerosol pollution, SST anomalies, and perturbations of the water cycle. The same is true for several additional though poorly quantified effects that aerosol particles can have on clouds and climate (Ramanathan et al. 2001). For example, the absorption of solar radiation by black carbon, which heats and thus stabilizes the aerosol pollution layer, could lead to the evaporation of clouds (Vogelman et al. 2001; Koren et al. 2004). Indirect aerosol effects on clouds also include the precipitation efficiency; for example, a high particle abundance may inhibit rainfall or suppress ‘warm’ rain formation in convective clouds (Rosenfeld and Woodley 2000; Rosenfeld et al. 2001). The latter effect can extend the vertical development of deep convective clouds, which promotes ice and hail formation and lightning so that some of these clouds may invigorate into heavy thunderstorms that produce torrential rain (Andreae et al. 2004; Koren et al. 2005; Chapter 18). To what degree these interactions between air pollution, clouds, and climate are relevant for the Mediterranean basin needs to be determined through coordinated research programmes. Global, regional, and local aspects influence both air pollution and climate, and mitigation or adaptation strategies should be based upon integrated problem assessments that also account for land use and soil hydrology changes. The largest risk lies in the possibility that some of these aspects combine

60°E

Fig. 22.10. Percentage changes in annual precipitation comparing low and high SST, calculated with a general circulation model based on observed SSTs (after Lelieveld et al. 2002b).

into destabilizing (positive) feedback mechanisms in the Mediterranean with potentially large consequences for a region that has been shown to be vulnerable to changes in air quality, climate, and the water cycle.

References Andreae, M. O., Rosenfeld, D., Artaxo, P., Costa, A. A., Frank, G. P., Longo, K. M., and Silvas- Dias, M. A. F. (2004), Smoking rain clouds over the Amazon. Science 303: 1337–42. Bardouki, H., Berresheim, H., Vrekoussis, M., Sciare, J., Kouvarkis, G., Oikonomou, K., Schneider, J., and Mihalopoulis, M. (2003), Gaseous (DMS, MSA, SO2 , H2 SO4 and DMSO) and particulate (sulphate and methanosulfate) sulfur species over the northeastern coast of Crete. Atmospheric Chemistry and Physic, 3: 1871–86. Barros, N., Toll, I., Soriano, R., Jiménez, P., Borrego, C., and Baldasano, J. M. (2003), Urban photochemical pollution in the Iberian Peninsula: Lisbon and Barcelona airsheds. Journal of the Air and Waste Management Association 53: 347–59. Berresheim, H., Plass-Dülmer, C., Elste, T., Mihalopoulos, N., and Rohrer, F. (2003), OH in the coastal boundary layer of Crete during MINOS: Measurements and relationship with ozone photolysis. Atmospheric Chemistry and Physics 3: 639–49. Bolle, H.-J. (2003), Climate, climate variability, and impacts in the Mediterranean area: an overview, in H.-J. Bolle (ed.), Mediterranean Climate—Variability and Trends. Springer, Berlin, 5–86. Corti, S., Molteni, F., and Palmer, T. N. (1999), Signature of recent climate change in frequencies of natural atmospheric circulation regimes. Nature 398: 799–802. Dayan, U., Heffter, J., Miller, J., and Gutman, G. (1991), Dust intrusion events into the Mediterranean basin. Journal of Applied Meteorology 30: 1185–98. de Castro, M., Gallardo, C., and Calabria, S. (2004), Regional IPCC projections until 2100 in the Mediterranean area, in

612

Jos Lelieveld

A. Marquina (ed.), Environmental Challenges in the Mediterranean 2000–2050. Kluwer, Dordrecht, 75–90. EMEP (2002), Emission data reported to UNECE/EMEP: Quality assurance and trend analysis and presentation of WebDab. MSCW Status Report by V. Vestreng and H. Klein. Meteorologisk Institutt: Norwegian Meteorological Institute Research Note 73 ISSN 0332-9879. Formenti, P., Andreae, M. O., Andreae, T., Galani, E., Vasaras, A., Zerefos, C., Amiridis, V., Orlovsky, L., Karnieli, A., Wendisch, M., Wex, H., Holben, B. N., Maenhaut, W., and Lelieveld, J. (2001), Aerosol optical properties and large-scale transport of air masses: Observations at a coastal and semiarid site in the eastern Mediterranean during summer 1998. Journal of Geophysical Research 106: 9807–26. Gangioti, M., Millán, M. M., Salvador, R., and Mantilla, E. (2001), Long-range transport and recirculation of pollutants in the western Mediterranean during the RECAPMA project. Atmospheric Environment 35: 6267–76. Ganor, E., Foner, H. A., Bingemer, H. G., Udisti, R., and Setter, I. (2000), Biogenic sulphate generation in the Mediterranean Sea and its contribution to the sulphate anomaly in the aerosol over Israel and in the Eastern Mediterranean. Atmospheric Environment 34: 3453–62. Henne, S., Furger, M., Nyeki, S., Steinbacher, M., Neininger, B., de Wekker, S. F. J., Dommen, J., Spichtinger, N., Stohl, A., and Prévôt, A. S. H. (2004), Quantification of topographic venting of boundary layer air to the free troposphere. Atmospheric Chemistry and Physics 4: 497–509. Husar, R. B., Prospero, J. M., and Stowe, L. L. (1997), Characterization of tropospheric aerosols over the oceans using the NOAA advanced very high resolution radiometer optical thickness operational product. Journal of Geophysical Research 102: 16899– 909. IPCC (Intergovernmental Panel on Climate Change) (2001), Climate Change 2001: The Scientific Basis. Cambridge University Press, Cambridge. Jiménez, P., Lelieveld, J., and Baldasano, J. M. (2006), Multiscale modeling of air pollutants dynamics in the northwestern Mediterranean basin during a typical summertime episode, Journal of Geophysical Research 111, D18306, doi: 10.1029/2005JD006516. Kallos, G., Kotroni, V., Lagouvardos, K., and Papadopoulos, A. (1998), On the long-range transport of air pollutants from Europe to Africa. Geophysical Research Letters 25: 619–22. Kentarchos, A. S., Roelofs, G.-J., and Lelieveld, J. (2001), Altitude distribution of tropospheric ozone over the Northern Hemisphere during 1996, simulated with a chemistry-general circulation model at two different horizontal resolutions. Journal of Geophysical Research 106: 17453–70. Kesselmeier, J. and Staudt, M. (1999), Biogenic volatile organic compounds (VOC): An overview on emission, physiology and ecology. Journal of Atmospheric Chemistry 33: 23–88. Koren, I., Kaufman, Y. J., Remer, L. A., and Martins, J. V. (2004), Measurement of the effect of Amazon smoke on inhibition of cloud formation. Science 303: 1342–5. Rosenfeld, D., Remer, L. A., and Rudich, Y. (2005), Aerosol invigoration and restructuring of Atlantic convective clouds. Geophysical Research Letters 32: L14828, doi: 10.1029/2005GL023187. Kormann, R., Fischer, H., de Reus, M., Lawrence, M., Brühl, C., von Kuhlmann, R., Holzinger, R., Williams, J., Lelieveld, J., Warneke, C., de Gouw, J., Heland, J., Ziereis, H., and Schlager,

H. (2003), Formaldehyde over the eastern Mediterranean during MINOS: Comparison of airborne in situ measurements with 3D-model results. Atmospheric Chemistry and Physics 3: 851–61. Kouvarakis, G. and Mihalopoulos, N. (2002), Seasonal variation of dimethylsulfide in the gas phase and of methanesulfonate and non-sea-salt sulfate in the aerosol phase measured in the eastern Mediterranean atmosphere. Atmospheric Environment 36: 929–938. Tsigaridis, K., Kanakidou, M., and Mihalopoulos, N. (2000), Temporal variations of surface regional background ozone over Crete island in the South-East Mediterranean. Journal of Geophysical Research 105: 4399–407. Vrekoussis, M., Mihalopoulos, N., Kourtidis, K., Rappenglueck, B., Gerasopoulos, E., and Zerefos, C. (2002a), Spatial and temporal variability of tropospheric ozone (O3 ) in the boundary layer above the Aegean Sea (eastern Mediterranean). Journal of Geophysical Research 107: 8137, doi: 10.1029/2000JD000081. Doukelis, Y., Mihalopoulos, N., Rapsomanikis, S., Sciare, J., and Blumthaler, M. (2002b), Chemical, physical and optical characterization of aerosol during the PAUR II experiment. Journal of Geophysical Research 107: 8141, doi: 10.1029/2000JD000291. Ladstätter-Weißenmayer, A., Heland, J., Kormann, R., von Kuhlmann, R., Lawrence, M. G., Meyer-Arnek, J., Richter, A., Wittrock, F., Ziereis H., and Burrows, J. P. (2003), Transport and tropospheric build-up of trace gases during the MINOS campaign: comparisons of GOME, in situ aircraft measurements and MATCH-MPIC-data. Atmospheric Chemistry and Physics 3: 1887–902. Lascaratos, A., Roether, W., Nittis, K., and Klein, B. (1999), Recent changes in deep water formation and spreading in the eastern Mediterranean Sea. Progress in Oceanography 44: 5–36. Lawrence, M. G., Rasch, P. J., von Kuhlmann, R., Williams, J., Fischer, H., de Reus, M., Lelieveld, J., Crutzen, P. J., Schultz, M., Stier, P., Huntrieser, H., Heland, J., Stohl, A., Forster, C., Elbern, H., Jakobs, H., and Dickerson, R. R. (2003), Global chemical weather forecasts for field campaign planning: predictions and observations of large- scale features during MINOS, CONTRACE, and INDOEX. Atmospheric Chemistry and Physics 3: 267–89. Lelieveld, J. and Dentener, F. J. (2000), What controls tropospheric ozone? Journal of Geophysical Research 105: 3531–51. and 30 others (2002a), Global air pollution crossroads over the Mediterranean. Science 298: 794–9. Peters, W., Dentener, F. J., and Krol, M. C. (2002b), Stability of tropospheric hydroxyl chemistry. Journal of Geophysical Research 107: 4715, doi: 10.1029/2002JD002272. Li, Q. B., Jacob, D. J., Bey, I., Palmer, P. I., Duncan, B. N., Field, B. D., Martin, R. V., Fiore, A. M., Yantosca, R. M., Parrish, D. D., Simmonds, P. G., and Oltmans, S. J. (2002), Transatlantic transport of pollution and its effects on surface ozone in Europe and North America. Journal of Geophysical Research 107, doi: 10.1029/2001JD001422. Lin, C.-Y., Jacob, D. J., and Fiore, A. M. (2001), Trends in exceedances of the ozone air quality standard in the continental United States. Atmospheric Environment 35: 3217–28. Long, M., Entekhabi, D., and Nicholson, S. E. (2000), Interannual variability in rainfall, water vapour flux, and vertical motion over West Africa. Journal of Climate 13: 3827–41.

Air Pollution and Climate Lövblad, G., Tarrasón, L., Tørseth, K., and Dutchak, S. (2004), EMEP Assessment Part 1: European perspective. Norwegian Meteorological Institute, Oslo (available from ). Luria, M., Peleg, M., Sharf, G., Siman Tov-Alper, D., Schpitz, N., Ben Ami, Y., Gawi, Z., Lifschitz, B., Yitzchaki, A., and Seter, I. (1996), Atmospheric sulfur over the East Mediterranean region. Journal of Geophysical Research 101: 25917–30. Markowicz, K. M., Flatau, P. J., Ramana, M. V., Crutzen, P. J., and Ramanathan, V. (2002), Absorbing Mediterranean aerosols lead to a large reduction in the solar radiation at the surface Geophysical Research Letters 29/20: 1968, doi: 10.1029/2002GL015767. Mihalopoulos, N., Stephanou, E., Kanakidou, M., Pilitsidis, S., and Bousquet, P. (1997), Tropospheric aerosol ionic composition above the Eastern Mediterranean Area. Tellus 49B: 314–26. Millán, M. M., Salvador, R., Mantilla, E., and Kallos, G. (1997), Photo-oxidant dynamics in the western Mediterranean in summer: Results from European research projects. Journal of Geophysical Research 102: 8811–23. Mantilla, E., Salvador, R., Carratalá, R., Sanz, M. J., Alonso, L., Gangioti, G., and Navazo, M. (2000), Ozone cycles in the western Mediterranean basin: Interpretation of monitoring data in complex coastal terrain. Journal of Applied Meteorology 39: 487– 508. Estrela, M. J., Sanz, M. J., Mantilla, E., Martin, M., Pastor, F., Salvador, R., Vallejo, R., Alonso, L., Gangioti, G., Ilardia, J. L., Navazo, M., Albizuri, A., Artiñano, B., Ciccioli, P., Kallos, G., Carvalho, R. A., Andrés, D., Hoff, A., Werhahn, J., Seufert G., and Versino, B. (2005), Climate feedbacks and desertification: The Mediterranean model. Journal of Climate 18: 684–701. Moulin, C., Lambert, C. E., Dulac, F., and Dayan, U. (1997), Control of atmospheric export of dust from North Africa by the North Atlantic Oscillation. Nature 387: 691–4. Mylona, S. (1996), Sulphur dioxide emissions in Europe 1880– 1991. Tellus 48B: 662–89. Nicholson, S., Some, B., and Kone, J. (2000), An analysis of recent rainfall conditions in ‘West Africa, including the rainy seasons of the 1997 El Niño and the 1998 La Niña years. Journal of Climate 13 2628–40. Nolle, M., Ellul, R., Heinrich, G., and Güsten, H. (2002), A longterm study of background ozone concentrations in the central Mediterranean—diurnal and seasonal variations on the island of Gozo. Atmospheric Environment 36: 1391–402. Ramanathan, V., Crutzen, P. J., Kiehl, J. T., and Rosenfeld, D. (2001), Atmosphere, aerosols, climate and the hydrological cycle. Science 294: 2119–24. Ribas, A. and Peñuelas, J. (2004), Temporal patterns of surface ozone levels in different habitats of the North Western Mediterranean basin. Atmospheric Environment 38: 985–92. Roelofs, G.-J., Scheeren, H. A., Heland, J., Ziereis, H., and Lelieveld, J. (2003), A model study of ozone in the eastern Mediterranean free troposphere during MINOS (August 2001). Atmospheric Chemistry and Physics 3: 1199–210. Rosenfeld, D. and Woodley, W. L. (2000), Deep convective clouds with sustained supercooled liquid water down to –37.5◦ C. Nature 405: 440–2. Rudich, Y., and Lahav, R. (2001), Desert dust suppressing precipitation: A possible desertification feedback loop. Proceedings of the National Academy of Sciences 98: 5975–80. Sateesh, S. K. and Ramanathan, V. (2000), Large differences in tropical aerosol forcing at the top of the atmosphere and earth’s surface. Nature 405: 60–3.

613

Schär, C., Vidale, P. L., Lüthi, D., Frei, C., Häberli, C., Liniger, M., and Appenzeller, C. (2004), The role of increasing temperature variability for European summer heat waves. Nature 427: 332–6. Scheeren, H. A., Lelieveld, J., de Gouw, J., van der Veen, C., and Fischer, H. (2002), Methyl chloride and other chlorocarbons in polluted air during INDOEX. Journal of Geophysical Research 107, doi: 10.1029/2001JD001121. Roelofs, G.-J., Williams, J., Fischer, H., de Reus, M., de Gouw, J. A., Warneke, C., Holzinger, R., Schlager, H., Klüpfel, T., Bolder, M., van der Veen, C., and Lawrence, M. (2003), The impact of monsoon outflow from India and southeast Asia in the upper troposphere over the eastern Mediterranean. Atmospheric Chemistry and Physics 3: 1589–608. Schneider, J., Borrmann, S., Wollny, A. G., Bläsner, M., Mihalopoulos, N., Oikonomou, K., Sciare, J., Teller, A., Levin Z., and Worsnop, D. R. (2004), Online mass spectrometric aerosol measurements during the MINOS campaign (Crete, August 2001). Atmospheric Chemistry and Physics 4: 65–80. Sciare, J., Cachier, H., Oikonomou, K., Ausset, P., Sarda-Estève R., and Mihalopoulos, N. (2003a) Characterization of carbonaceous aerosols during the MINOS campaign in Crete, July–August 2001: a multi-analytical approach. Atmospheric Chemistry and Physics 3: 1743–57. Bardouki, H., Moulin, C., and Mihalopoulos, N. (2003b), Aerosol sources and their contribution to the chemical composition of aerosols in the Eastern Mediterranean Sea during summertime. Atmospheric Chemistry and Physics 3: 291–302. Simon, V., Dumergues, L., Solignac, G., and Torres, L. (2005), Biogenic emissions from Pinus halapensis: a typical species of the Mediterranean area. Atmospheric Research 74: 37–48. Stott, P. A., Stone, D. A., and Allen, M. R. (2004), Human contribution to the European heat wave of 2003. Nature 432: 610–14. Tragou, E. and Lascaratos, A. (2003), Role of aerosols on the Mediterranean solar radiation. Journal of Geophysical Research 108: 3025, doi:10.1029/2001JC001258. Traub, M. and Lelieveld, J. (2003), Cross-tropopause transport over the eastern Mediterranean. Journal of Geophysical Research 108: 4712, doi: 10.1029/2003JD003754. Fischer, H., de Reus, M., Kormann, R., Heland, J., Ziereis, H., Schlager, H., Holzinger, R., Williams, J., Warneke, C., de Gouw, J., and Lelieveld, J. (2003), Chemical characteristics assigned to trajectory clusters during the MINOS campaign. Atmospheric Chemistry and Physics 3: 459–68. van Aardenne, J. A., Dentener, F. J., Olivier, G. J., Klein Goldewijk, C. G. M., and Lelieveld, J. (2001), A 1◦ × 1◦ resolution data set of historical anthropogenic trace gas emissions for the period 1890–1990. Global Biogeochemical Cycles 15: 909–28. Vestreng, V., Breivik, K., Adams, M., Wagner, A., Goodwin, J., Rozovskaya, O., and Pacyna, J. M. (2004), Inventory review, emission data reported to CLRTAP and under the NEC Directive, EMEP/EEA Joint Review Report, EMEP/MSC-W Note 1/2004. Vogelman, A. M., Ramanathan, V., and Podgorny, I. A. (2001), Scale dependence of solar heating rates in convective cloud systems with implications to general circulation models. Journal of Climate 14: 1738–52.

614

Jos Lelieveld

Vrekoussis, M., Kanakidou, M., Mihalopoulos, N., Crutzen, P. J., Lelieveld, J., Perner, D., Berresheim, H., and Baboukas, E. (2004), Role of the NO3 radicals in oxidation processes in the eastern Mediterranean troposphere during the MINOS campaign. Atmospheric Chemistry and Physics 4: 169–82.

Ward, M. (1992), Provisionally corrected surface wind data, worldwide ocean-atmosphere surface fields and Sahelian rainfall variability. Journal of Climate 5: 454–75. Zerefos, C., Ganev, K., Kourtidis, K., Tzortsiou, M., Vasaras, A., and Syrakov, E. (2000), On the origin of SO2 above Northern Greece. Geophysical Research Letters 27: 365–8.

This chapter should be cited as follows Lelieveld, J. (2009) Air pollution and climate, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 599–614.

23

Biodiversity and Conservation Jacques Blondel and Frédéric Médail

Introduction The biodiversity of Mediterranean-climate ecosystems is of particular interest and concern, not only because all five of these regions (the Mediterranean basin, California, central Chile, Cape Province of South Africa, western and southern parts of Australia) are among the thirty-four hotspots of species diversity in the world (Mittermeier et al. 2004), but they are also hotspots of human population density and growth (Cincotta and Engelman 2000). This relationship is not surprising because there is often a correlation between the biodiversity of natural systems and the abundance of people (Araùjo 2003; Médail and Diadema 2006) and this, inevitably, raises conservation problems. Within the larger hotspot of the Mediterranean basin as a whole, ten regional hotspots have been identified (Figure 23.1). They cover about 22 per cent of the basin’s total area and harbour about 44 per cent of Mediterranean endemic plant species (Médail and Quézel 1997, 1999), as well as a large number of rare and endemic animals (Blondel and Aronson 1999). A key feature of these Mediterranean hotspots as a whole is their extraordinarily high topographic diversity with many mountainous and insular areas. Not surprisingly this results in high endemism rates and they contain more than 10 per cent of the total plant richness (see the recent synthesis of Thompson 2005). However, of all the mediterranean-type regions in the world, the Mediterranean basin harbours the lowest percentage (c.5%) of natural vegetation considered to be in ‘pristine condition’ (Médail and Myers 2004; Chapter 7). With an average of as many as 111 people per km2 , one may expect a significant decline in biological diversity in the Mediterranean basin—a region that has been managed, modified, and, in places, heavily degraded by humans for millennia (Thirgood 1981; Braudel

1986; McNeill 1992; Blondel and Aronson 1999; Chapter 9). There are two contrasting theories that consider the relationships between humans and ecosystems in the Mediterranean (Blondel 2006, 2008). The first one is the ‘Ruined Landscape or Lost Eden’ theory, first advocated by painters, poets, and historians in the sixteenth and seventeenth centuries, and later by a large number of ecologists. This theory argues that human action in the form of deforestation and overgrazing resulted in a progressive and cumulative degradation and desertification of Mediterranean landscapes. This theory denounces the destruction of the formerly extensive forests which were supposedly so lush and large that a monkey could have travelled from Spain to Turkey almost without leaving the crown of the trees! One example of this view has been vividly depicted by David Attenborough in his book, The First Eden: The Mediterranean World and Man (Attenborough 1987; see also Naveh and Dan 1973; Thirgood 1981; McNeill 1992). Challenging this pessimistic view, the second school dismisses the supposedly detrimental effects of humans by arguing that the imaginary past idealized by artists and scientists does not acknowledge the fact that humans actually contributed to the maintenance of Mediterranean landscapes as they were progressively established since the last glacial episode, stressing that savanna-like landscapes are characteristic of the Mediterranean (Grove and Rackham 2001). Reality is inevitably somewhere between these two extremes which are of crucial relevance to any discussion of threats to biodiversity in this region. In fact, several modelling attempts suggest that the Mediterranean region is one of the most threatened regions in the world with regards to environmental changes (Sala et al. 2000). Generally speaking, all the components of

616

Jacques Blondel and Frédéric Médail

Maritime and Ligurian Alps

Western Mediterranean Islands Betic–Rifan complex High and Middle Atlas

Djurdjura– Kabylies

Southern and Central Greece Crete

Southern Anatolia and Cyprus

Syria, Lebanon and Israel

Mediterranean Cyrenaica

0

500 km

Fig. 23.1. The ten regional species diversity hotspots of the Mediterranean basin based on plant endemism and richness (based on Médail and Quézel 1997, 1999 and Véla and Benhouhou 2007).

‘global change’ (Vitousek 1994) threaten, to various extents, the biodiversity in this region, but the component of most concern is human population growth and its consequences for species, habitats, and ecosystems. Indeed, the resident population has increased from 213 million people in 1950 to 427 million in 2000, with 523 million expected in 2025 and 600 million in 2050, mainly located along the coasts (Benoit and Comeau 2005). Rather than reviewing all the threats that currently affect Mediterranean biotas, this chapter will focus on some of the most important drivers—such as habitat destruction and fragmentation, and the impacts of fire and invasive species—and try to make some generalizations about habitat dynamics and the impact of these threats upon the most important and/or wellknown groups of terrestrial and freshwater plants and animals.

Contrasted Human Impact in Space and Time What characterizes ecosystems and habitats much more in the Mediterranean region than in any other region in the world is their long-lasting common history with

humans as they have been designed and redesigned by them for almost 10,000 years in the eastern part of the basin and around 8,000 years in its western part (e.g. Le Houérou 1981; Braudel 1986; Pons and Quézel 1985). However, the action of humans has been far from always detrimental and has sometimes resulted in an increase in biodiversity through the shaping of a large variety of cultural landscapes. For example, Blondel and Aronson (1995, 1999) argue that many traditional land use practices act as surrogates of natural disturbance regimes with the consequence that, according to the intermediate disturbance hypothesis (Huston 1994), several components of biodiversity have actually been much higher in landscapes shaped by humans than in primitive plant communities such as oak woodlands. The coexistence of stock-farming, crop fields, and wildlife has led to higher local diversity in many parts of the Mediterranean so that abandoning sustainable and moderate land use practices inevitably decreases the biodiversity of many groups. Examples of traditional agrosylvo-pastoral land use systems include the Roman triad known as sylva-saltus-ager (woodland-pasture-field) in which landscapes are divided into distinct plots devoted to wood cutting (sylva), livestock breeding (saltus), or crop harvesting (ager). Another traditional system is the

Biodiversity and Conservation

617

Fig. 23.2. A montado in Portugal with cattle under the trees and charcoal burners (photo: Jacques Blondel).

Montado-dehesa system of the Iberian Peninsula and several Mediterranean islands where the three activities mentioned above occur in the same plots (Figure 23.2). Thanks to the fine balance of woodlots, pastoral grasslands, shrublands, and open spaces reserved for cultivation, the resulting ‘moving mosaic’ greatly contributed to the biological diversity of Mediterranean landscapes (Blondel and Aronson 1995, 1999; Grove and Rackham 2001). Although many of these ancient land use practices have almost disappeared in most parts of the basin, their imprint on the structure of landscapes and habitats is still visible today. This is an illustration of the value for biodiversity of cultural landscapes and habitats and the evolving history of land use impacts by humans (Foster 2002). Over past millennia and increasingly as cultural impacts became more varied, ecosystems and habitats have changed continuously, supporting mosaics of highly dynamic landscapes (Blondel and Aronson 1999; Foster et al. 1990). Most of them have been influenced and shaped by cultural activity for so long that it is

often difficult to decipher the natural from the human components of their history. The impact of human land use and disturbance legacies is apparent in all ecosystems (Foster 2002). For example, in an archaeological study of French forests, Dupouey et al. (2002) demonstrated that the signature of agriculture and pastoralism dating back to the Romans is still visible today in the form of a mosaic-like distribution of certain plant species (e.g. Vinca minor, Ribes uva-crispa) which are typically linked to ancient human settlements. Furthermore, some recent palaeoecological studies suggest that the forest cover was not as dense and uniform as formerly thought during the Holocene optimum, with many landscapes being rather open and heterogeneous, but also with some strong regional differences in vegetation dynamics (Beaulieu et al. 2005). However, the idyllic vision of a ‘natural sustainability’ driven by traditional agro-pastoral systems is not easy to sustain as shown by the many examples of ancient and devastating human impacts on Mediterranean ecosystems (Thirgood 1981; Chapter 9).

618

Jacques Blondel and Frédéric Médail

The present trends of habitat changes as a result of human impact vary markedly across the regions of the Mediterranean basin (Barbero et al. 1990; Mazzoleni et al. 2004; Blondel 2008). In the northern parts, the collapse of traditional land use systems and rural depopulation from the end of the nineteenth century, and especially after the two world wars, completely destructured traditional landscapes as they were replaced by industrial-style agriculture and modern activities such as mass tourism. Since the beginning of the twentieth century, changes in land use practices in the European parts of the Mediterranean included industrial-scale cultivation, but also a rapid transformation of previously cultivated or grazed lands towards shrubby and forested areas, leading to what may be described as a homogenization of the flora and fauna. As a result, thermophilous taxa of Mediterranean origin tend to be replaced by more ubiquitous species of medioEuropean origin. One example is the replacement of Mediterranean warblers—which are typical of Mediterranean shrubby habitats—such as the Sardinian warbler Sylvia melanocephala or the Spectacled warbler Sylvia conspicillata by forest species such as the Blackbird Turdus merula or the Blue tit Cyanistes caeruleus (Blondel and Farré 1988; Debussche et al. 2001). In the northern, European side of the Mediterranean, human pressure on many ecosystems has steadily decreased as a result of agricultural abandonment and rural depopulation dating back to the end of the nineteenth century, which has accelerated greatly since the Second World War. Across the entire range of life zones and habitat types, a progressive recovery of forests and matorrals invading old fields, and abandoned terraces and vineyards is evident (Mazzoleni et al. 2004; Figure 23.3; Chapter 7). These processes lead to a decrease in habitat patchiness and a slowing down of the ‘moving mosaic’, that is the turnover in time of habitat patches according to how farmers manage them, which is so characteristic of Mediterranean landscapes and beneficial for several components of biological diversity. The collapse of traditional land use systems that maintained habitat heterogeneity during centuries undoubtedly had and will have consequences on biodiversity and its distribution at the landscape scale. In contrast, in North Africa and most of the eastern part of the Mediterranean, pressures on natural habitats by human populations and livestock are still strongly on the increase, destructuring soils and ecosystems and resulting in intense erosion. Woodward (1995) mentioned that the impact of ongoing woodland clearance followed by overgrazing has resulted in high rates of soil erosion across the Mediterranean drainage basin.

The average area annually cultivated with cereals in North Africa increased by 50 per cent during 1948– 52 and 1981–90 (Le Houérou 1990). This resulted in more extensive clearing of arid steppe rangeland which is a major cause of desertification in North Africa and the Near East. As a result, disturbance regimes in farm, pasture, and forest lands are moving towards still greater intensity of land and resource exploitation for the shortterm survival of local people. With an average population growth of 3.2 per cent per year in North Africa and several countries of the Near East, the consequences of increased human pressure in the arid zone of the southern shores of the Mediterranean could be extensive runoff, frequent flooding, and extreme degradation of the remaining patches of forest (Figure 23.4). Soil erosion, which averages 5–10 t ha−1 yr−1 in medium to large catchments, might increase by a factor of five or more to reach average values of 25–50 t ha−1 yr−1 as measured in limited areas of Algeria and Morocco (ibid.) where 35 per cent of land is already undergoing losses exceeding 30 t ha−1 yr−1 (see Chapter 20). These opposite trends that characterize the dynamics of ecosystems on the two sides of the Mediterranean have markedly contrasting impacts on the wealth of biological systems with a trend of habitat regeneration in the north and a trend of ongoing degradation in the south. In the northern part of the basin, forest is recovering at a rate of 2 per cent per year as a result of rural depopulation and land abandonment, whereas it is almost exactly the opposite in the eastern and southern parts of the basin (Quézel and Médail 2003).

Habitat Destruction All habitats and landscapes in the Mediterranean region, except perhaps some remote mountainous and steep cliff areas, have been to some extent managed and transformed by humans. These tremendous human-induced changes have had profound consequences on biodiversity. Although some of them have had beneficial consequences on the dynamics and distribution of species, communities, and habitats as mentioned above, many have resulted in serious threats to biodiversity. Forests, wetlands, and coastal habitats, together with their associated flora and fauna, are perhaps those that have been, and continue to be, most affected by deforestation, land reclamation, and human encroachment on coastlines (Ramade 1997). A gap analysis suggests than no more than 5–10 per cent at most of the native postglacial Mediterranean forest is left (WWF 2001) with huge geographical

Biodiversity and Conservation

619

Fig. 23.3. Forest recovery on ancient terraces (photo: Jacques Blondel).

variation. The main threats to forests are mismanagement associated with overexploitation of resources such as timber, firewood, and grazing—especially in the southern and eastern parts of the basin (Quézel and Médail 2003). A particularly dramatic situation is

that of the forests or forest-steppes of the Maghreb with native woodland dominated by Cedrus atlantica, Quercus canariensis, Quercus faginea, or Juniperus thurifera being reduced to half of their native extent or seriously disturbed, and the former beautiful forests of Tetraclinis

620

Jacques Blondel and Frédéric Médail

Fig. 23.4. The last individual of a former forest of Juniperus thurifera in southern Morocco (photo: Frédéric Médail).

Biodiversity and Conservation

articulata have almost disappeared (Quézel 2000). The dramatic decline of the famous cork oak forest of the Mamora in northern Morocco in only seventy-five years is a striking example of the magnitude of forest destruction in North Africa. This forest extended over 130,000 ha in 1920 but was reduced to 77,064 ha in 1960, 60,000 ha in 1980, and a mere 12,844 ha in 1995. The degradation and draining of wetlands is not a recent problem but has a history going back several centuries during which many coastal wetlands have disappeared (Chapter 9). For example, the Etruscans began to drain the marshes around Rome as early as the fifth century AD (Pearce and Crivelli 1994). Most of the 250 freshwater or brackish lakes and marshes of Central Anatolia (180 of which are more than 100 ha in size) have already disappeared, many of them having been drained in the 1950s and 1960s in order to tackle malaria and to develop new agricultural land. Between 1984 and 2002, the surface area of the Anatolian marshes of the Konya plain was reduced by 90 per cent and the most striking case is that of the 10,000 ha of the Yarma marshes which were completely dried up within ten years (Gramond 2002). Many wetlands that served as refugia during glacial times for Euro-Siberian and boreal plant species (e.g. Andrieu-Ponel et al. 2000) and which were initially designated by Roi (1937) as colonies planitiaires (e.g. the lower Rhône in southern France, the Arno and Serchio estuaries in Tuscany, and the Pontin marshes near Rome) have been drained, sending many local populations of relict flora and fauna to extinction. In the early 1920s, Greek wetlands covered three times their present area, and a third of the remaining wetlands are threatened. In Macedonia alone, 1,151 km2 of wetlands out of a total of 1,572 km2 have been drained since 1930 (Catsadorakis 2003). Because wetlands are very productive ecosystems, they are also much desired by humans and many of them are managed for the production of various crops. For example, the delta of the River Axios in Greece is increasingly threatened by the expansion of the large city of Thessaloniki with mussel beds and ricefields providing 90 and 70 per cent respectively of the total Greek production. In spite of these encroachments, the delta is still an important breeding and stopover place for 250 species of birds, 76 of which are rare, and 36 species of fish. Even in prestigious and well-protected areas such as the Camargue, habitat losses have not stopped since the nineteenth century with a net loss of c.40,000 ha during the second half of the twentieth century (Tamisier and Grillas 1994; Figure 23.5).

621

The high consumption of water for agriculture and tourism induces an over-exploitation of groundwater around the Mediterranean, especially in Israel, Spain, Malta, Cyprus, Tunisia, and southern Italy, contributing to extensive drainage of wetlands. Water consumption peaks during the summer, when the resources are at their lowest, with each Mediterranean tourist using between 300 and 850 litres of water per day (De Stefano 2004). The result of this tremendous increase in water demand has been the destruction of 1 million hectares of wetlands within the Mediterranean region over the last fifty years (Skinner and Zalewski 1995; Chapter 21). Riparian forests are associated with wetlands and they share a great many plant and animal species. These environments are well known for their diversity as well as for their role in controlling floods. In addition, they act as natural corridors allowing many species of temperate Europe to colonize Mediterranean habitats, greatly contributing to their biological diversity (Décamps and Décamps 2002). Unfortunately most of the Mediterranean riparian forests have been destroyed and replaced by intensive agriculture or by stands of trees such as poplars. Among other aquatic habitats of particular interest in the Mediterranean are the temporary and oligotrophic ponds, which constitute species-rich and original habitats with several rare and endangered ferns (Pilularia, Isoëtes, Marsilea), phanerogams (Damasonium), insects, frogs, and large branchiopods such as Triops cancriformis (Grillas et al. 2004; Figure 23.6). They are also key breeding sites for amphibians, including the rare Mediterranean newt Triturus marmoratus (Figure 23.7). Their characteristic species are adapted to the drastic and unpredictable conditions imposed by summer aridity, interannual fluctuations in hydrology, and a poor supply of nutrients. These temporary marshes are indeed of major conservation importance, but they have been considerably degraded by various disturbance events, especially in North Africa (Rhazi et al. 2001; Grillas et al. 2004). Such environments will also be threatened by climate change given the scenarios associated with a warmer climate in the region (Chapter 3). Coastal habitats and ecosystems are also increasingly under threat. Rapid changes in land use practices in the twentieth century, especially over the last four decades, have had disastrous consequences for coastal ecosystems where more than 60 per cent of people live. In south-eastern France, for example, population density reaches 2,500 people per km2 in the Alpes-Maritimes département compared with an average of just 108 people per km2 in the country as a whole (Figure 23.8).

622

Jacques Blondel and Frédéric Médail

Fig. 23.5. Wetlands are among the most threatened habitats in the Mediterranean region (photo: Jacques Blondel).

In Italy, 70 per cent of the coast is already urbanized and rates of human encroachment on coastal areas are growing at an unprecedented rate in many Mediterranean regions such as southern Spain, the coasts of the Adriatic Sea, and many countries of the Near East. Rampant urban development and pollution are serious threats to coastal habitats because in many areas urban waste is pumped untreated directly into the sea. Forming a zone where terrestrial and aquatic habitats meet, the coast is especially fragile and vulnerable. For example, 60 per cent of the Greek population, 40 per cent of Greek agriculture, 70–80 per cent of Greek industrial activity, and 90 per cent of tourism concentrate in the coastal zone (Catsadorakis 2003). The accelerated rate of urbanization of coastlines across the Mediterranean is likely to reach 75–80 per cent almost everywhere in the basin by 2025. The high demand for Mediterranean coastal and insular landscapes—because of their beauty and popularity with tourists—makes them particularly threatened (Delanoë et al. 1996), especially because of their well-known vulnerability owing to a high pro-

portion of rare endemic species and unique insular habitats.

Habitat Fragmentation Landscape fragmentation is a very ancient practice everywhere around the Mediterranean, adding to the natural diversity of habitats, especially along the coasts and around urban areas. If fragmentation may threaten biodiversity by increasing inbreeding and extinction risks for small isolated populations (Noss and Csuti 1994), the process is not necessarily detrimental overall, as shown by the consequences of traditional land use practices on several components of biodiversity such as alpha, beta, and gamma diversity (see Blondel and Aronson 1995, 1999). For example, the fragmentation of oak forests in southern France resulted in a local increase of species richness of many groups, especially ruderal plant species that benefit from edge effects along roads and crop fields where nitrogen-rich nutrients are imported by humans. However, the richness of different plant functional types in the core areas of forests

Biodiversity and Conservation

623

Fig. 23.6. The crustacean branchiopod Triops cancriformis with the rare thrumwort Damasonium stellatum and the parsley frog Pelodytes punctatus in a temporary marsh in the Camargue (Blondel and Aronson 1999).

remains higher or equal to that of the edges (Médail et al. 1998). Fragmentation may seriously change the dynamics of biological interactions. For example, as a result of high densities of seed eaters (‘supersaturation’) such as voles in habitat mosaics resulting from fragmentation, seed recruitment of forest species may be modified. In Mediterranean Spain, the average number of seedlings of the holm oak Quercus ilex is nine times higher in relatively large undisturbed forests (150–350 ha) than in small forest fragments (0.2–12 ha in size) where most acorns are eaten by animals (Santos and Tellería 1997). Post-dispersal loss may also be exacerbated within small isolated habitats. For example, up to 25 per cent of the fruits of Juniperus thurifera may be eaten by voles (Apodemus sylvaticus) in small habitat fragments, whereas such losses do not exceed 5 per cent in large forest tracts (Santos and Tellería 1994).

Habitat fragmentation may also have serious consequences on the fitness, genetic diversity, and local adaptation of organisms. For example, in Cyclamen balearicum, a rare endemic plant species from southern France and the Balearic Islands, isolation and fragmentation of sclerophyllous oakwoods resulted in a loss of genetic diversity, an increase of genetic drift, and selection-biased mechanisms leading to spontaneous self-pollination in highly isolated populations (Affre et al. 1997; Thompson 2005). A long-term programme has investigated the causes and consequences of habitat fragmentation on phenotypic plasticity (that is, the production by the same genotype of different phenotypes depending on local habitat features) in a song bird, the Blue tit (Cyanistes caeruleus). Results showed that depending on the size of patches within habitat mosaics and the type of oak, whether deciduous (Quercus humilis) or evergreen (Q. ilex), this

624

Jacques Blondel and Frédéric Médail

Fig. 23.7. A rare amphibian of temporary ponds of the western part of the basin, the Mediterranean newt, Triturus marmoratus (photo: N. Zbinden).

small passerine responds in several ways to habitat fragmentation. In a mainland landscape in southern France, where birds are assumed to disperse freely across habitat patches, little local differentiation of two breeding traits—laying date and clutch size—has been found within a range of c.40 km, suggesting gene swamping between populations (Blondel et al. 2001; Figure 23.9). A molecular genetics study of populations within this landscape supported the hypothesis of a source-sink population structure with more birds immigrating from deciduous habitat patches to evergreen ones than the reverse (Dias et al. 1996). In a similar geographic configuration of habitats in Corsica, there was a much higher phenotypic variation and a higher degree of population differentiation on a scale that is usually smaller than the dispersal range of Blue tits. This difference between mainland France and Corsica has been interpreted as resulting from reduced dispersal ranges of birds on islands and supports the divergence-with-gene-flow model of speciation (Blondel et al. 1999; Blondel 2008). In the mainland landscape where mismatching between breeding time and the peak of food abundance is high, maladaptation can result in poor breeding success making such populations dependent on source populations at the scale of landscapes. This implies that conserva-

tion strategies must consider the geographical configuration of habitats at large spatial scales. The responses of Blue tits to habitat patchiness show how organisms can respond to environmental changes including climate change. The changes that happened many times during the Pleistocene are currently occurring at an accelerated speed as a result of global warming. This may create problems for organisms across the Mediterranean region as they attempt to adapt (Jump and Peñuelas 2005). All the empirical evidence suggests that the biological consequences of habitat fragmentation should be carefully studied, because they are complex and generally underestimated.

Is Fire a Real Threat to Mediterranean Biodiversity? On average 1 per cent of the Mediterranean forests, that is c.600,000 ha, burn annually with many consequences for human societies. Wildfires frighten people and have detrimental socio-economic effects, but provided their frequency does not exceed a certain threshold they are major components of the dynamics of Mediterranean-type ecosystems and do not necessarily

Biodiversity and Conservation

625

Fig. 23.8. The urbanization of coastal areas threatens habitats and rare plants and animals almost everywhere in the Mediterranean basin. On this photo the city of Monaco is built on a habitat for the rare woody shrub Euphorbia dendroides that forms the bright rounded clumps in the foreground (photo: Frédéric Médail).

threaten biodiversity (Moreno and Oechel 1994; Eugenio and Lloret 2006; Chapter 19). The use of fire by humans is a very old practice in the Mediterranean region, perhaps as ancient as 700,000 BP in northern Israel (Goren-Inbar et al. 2004). However, regular use of fire became more frequent only between c.7,000 and 5,000 BP, mainly to clear vegetation for agricultural and pastoral purposes, but also as a strategic weapon in the wars of Antiquity (KuhnholtzLordat 1938; Pons and Thinon 1987). In modern times, c.50,000 wildfires occur annually, most of them caused by humans—only 5 per cent at most are natural fires generated by lightning. The majority of burnt hectarage is caused by a few very large wildfires: in coastal eastern Spain, 144 large fires (i.e. only 1.3% of the total) destroyed 487,793 ha (i.e. 78% of the total burnt surface areas) between 1968 and 1994 (Piñol et al. 1998; Chapter 19). From a landscape perspective, Mediterranean fires are disturbance events that maintain habitat heterogeneity

and the moving mosaic that keeps ecosystems functioning. For example, Naveh (1999) demonstrated, in his studies in the Mount Carmel area of Israel, the key role of fire as an evolutionary and ecological factor in shaping landscapes and vegetation. Many communities and species of open habitats, including xero-thermophilous species, are narrowly dependent on such disturbance events. One example of the role of fire in maintaining bird communities in Mediterranean landscapes has been provided by Prodon et al. (1987). At the scale of the whole range of habitats within a landscape consisting of a habitat mosaic generated by recurrent fire events in southern France, fifty-one species of birds have been found, each of them occupying its preferred habitat. Bird species of open vegetation such as larks (e.g. Lullula arborea), partridges (e.g. Alectoris rufa), and linnets (Carduelis cannabina) colonize habitats immediately after fire and are then replaced by species of matorrals such as warblers (Sylvia spp.) and the Nightingale (Luscinia megarhynchos) as habitats recover. In turn, matorral

626

Jacques Blondel and Frédéric Médail

Laying date (1=1st of March)

100

80

CE1

CE3 Evergreen

CD1 ME1

CE2

60

ME2 CD2

40

MD1

ME3

MD2 Deciduous

20 Corsica

Mainland

Fig. 23.9. Mean laying date (in ‘March-date’, i.e. 32 = 1 April) of the Blue tit Cyanistes caeruleus in mainland and Corsican habitats. The horizontal dotted lines indicate the best date for onset of breeding relative to resource availability (leaf-eating caterpillars) in deciduous habitats (D) and in evergreen habitats (E), respectively. CE = ‘Corsica evergreen’, CD = ‘Corsica deciduous’, ME = ‘Mainland evergreen’, MD = ‘Mainland deciduous’. Vertical bars = 1 standard deviation (modified from Blondel et al. 2001).

species will be replaced by species of later transitional stages and so forth until the recovery process is completed, returning to a forest stage with such species as nuthatches and woodpeckers. Fire events periodically move up and down the position of any given habitat patch on the grassland → matorral → forest system. At a broader geographical scale, the combination of a disturbance regime and community dynamics, resilience, and inertia results in an equilibrium that may be fairly stable in the long term (Blondel 1987). Communities of small mammals have been shown to recover quickly after fire with a definite sequence of local colonizations and extinctions that match successional changes in the structure of vegetation (Prodon et al. 1987; Torre and Diaz 2004). Several species of small mammals, especially the wood mouse Apodemus sylvaticus and to a lesser extent the white-toothed shrew Crocidura russula and the Algerian mouse Mus spretus, have been shown to be more abundant in the early post-fire habitats than later on in transitional stages of the post-fire habitat gradient, presumably as a result of the combination of large quantities of seeds and seedlings in these herbaceous areas, as well as alleviated predation pressures in recently burnt areas (Torre and Diaz 2004). An amazing and unexpected fact in burnt habitats is the local persistence after fire of Mediterranean land snail communities (Kiss et al. 2004; Kiss and Magnin 2006), presumably because of the availability of underground microrefuges.

The post-fire resilience of Mediterranean vegetation is a well-known characteristic that is due to several intrinsic properties of plants (Trabaud and Prodon 2002; Eugenio and Lloret 2006). These include the resprouting capacities of plants with the existence of lignotuber, the role of the edaphic seed-bank, and the colonization of burnt areas by seed dispersal through wind or vertebrates (Quézel and Médail 2003). Only very high fire return rates can induce a real malfunctioning of ecological processes resulting in threats to biodiversity and increased runoff and soil erosion. Repeated fires combined with soil erosion have caused drastic edaphic changes in some postglacial soils and have contributed to a strong decline in Mediterranean old-growth forests formed in part by laurifolious (e.g. Ilex aquifolium, Taxus baccata, Laurus nobilis) and deciduous (e.g. Quercus spp., Celtis australis, Cercis siliquastrum) trees. One of the consequences of erosion and ecosystem degradation on limestone soils is the so-called leopard’s skin effect, with an alternation of bare ground and stones with low bushes of dwarf oaks (Quercus coccifera) or spiny xerophytes (e.g. Astragalus, Sarcopoterium, Ulex) as shown in Figure 23.10. To the authors’ knowledge, however, no plant species has become extinct as a result of fire. In fact, some rare and threatened plants are actually favoured by fire events. Examples in southern France include several species of Fabaceae (e.g. Genista linifolia, Vicia altissima, Vicia melanops) whose seed germination is enhanced by high temperatures, which is also the case for most of the Mediterranean geophytes (e.g. Acis, Iris, Orchis, Serapias, Tulipa). A detailed study of a rare endemic geophyte of the Maritime Alps (Acis nicaeensis, Amaryllidaceae) indicated that, a year after the event, the fire induces a significant increase in the density of flowering individuals and seedling emergence, as well as in clump densities, by reducing aboveground plant competition and increasing bare soil cover (Diadema et al. 2007). However, such differences were quickly attenuated two years after the fire event. These results suggest that small-scale fires can be beneficial for the regeneration window of this threatened geophyte, through the periodic supply of the seed and bulb banks in the soil.

Invasive Species Of all the threats to biodiversity in the Mediterranean, the introduction of alien species as a side-effect of human colonization has been particularly harmful for many biotas. The introduction of exotic species can cause considerable disruption to native ecosystems (e.g. Drake

Biodiversity and Conservation

627

Fig. 23.10. Intense habitat degradation due to repeated fire events results in very low scrubby vegetation and bare ground, the so-called leopard’s skin (photo: Jacques Blondel).

et al. 1989; Williamson 1996) so that biological invasions are currently considered as one of the main threats to biodiversity (Mooney and Hobbs 2000). In some regions of the world, especially on islands, invading species have displaced whole cohorts of native species and disrupted ecosystem dynamics; thus they are increasingly considered potential threats to biodiversity and ecosystem functioning. However, this assertion was recently questioned by Gurevitch and Padilla (2004) who argue that on a worldwide scale endangered taxa are more than five times more threatened by habitat loss than by biological invasions, with alien species contributing less than 2 per cent to the extinction of the 762 documented plants and animals in the world IUCN (International Union for Conservation of Nature) database. Thus, the primary role of invasive aliens in driving widespread extinctions has not yet been clearly demonstrated; more research is needed to estimate the real threats more accurately. On the whole, the Mediterranean basin appears to be less vulnerable to species invasion than the other

mediterranean-type regions of the world, especially California and the Cape Province of South Africa. It has been suggested that the low invasiveness of mediterranean ecosystems, except for some habitats such as freshwaters, wetlands, or riparian forests, results from the strong interactions that tied humans and ecosystems for millennia, with recurrent human-induced disturbance regimes, resulting in some kind of coevolution between them (di Castri et al. 1990). Combined with thousands of spontaneous colonization events and selection pressures by humans, disturbance regimes made ecosystems progressively more resistant with ‘old invaders’ preventing potential ‘new invaders’ from colonizing them (Drake et al. 1989). Among fish species, ancient introduced species such as the carp (Cyprinus carpio) or the perch (Perca fluviatilis) can be considered as naturalized, but many recently introduced species pose real threats for the native fauna. Invading taxa may account for more than half the species in some water bodies. For example, nineteen species in Spain as compared to thirty-two

628

Jacques Blondel and Frédéric Médail

native, and ten species in Portugal as compared to twenty-nine native have been introduced (Crivelli and Maitland 1995). The situation seems to be better in some other regions such as Greece where only 11 of the 117 fish species of the country have been introduced. Some species are particularly harmful for the indigenous fish fauna, including predators such as the sheatfish (Silurus glanis) the pike-perch (Sander lucioperca), and the bass (Micropterus salmoides). The introduction of herbivorous carp from China in 1983 in Lake Oubeira, Algeria, resulted in the destruction of half the reed-beds of this marsh (Pearce and Crivelli 1994) with disastrous consequences for the native biotas. A severe drought completely desiccated this lake some years ago, however, killing all fish species including the introduced carp. The Louisiana crayfish (Procambarus clarkii) deserves special attention—it invaded many rivers and wetlands of the region, causing irreversible damage to many communities of invertebrates. Exotic crayfish species are an example of the paradox of invasions because they may benefit populations of emblematic and popular predators such as herons that feed on them. Unfortunately the negative side is an undetected and irreversible damage to other components of communities that make ecosystems function. Relatively few exotic birds and mammals have invaded Mediterranean ecosystems. Even the Pheasant (Phasianus colchicus), which was introduced by the Romans, did not succeed in establishing self-sustaining free populations and the many attempts to introduce species as game birds (the Californian quail Colinus virginianus, the Black francolin Francolinus francolinus, and some others) usually met with failure. The only species that established themselves in the wild are small populations of the Ring-necked parakeet Psittacula krameri in several large cities around the Mediterranean, free-ranging populations of the Red avadavar Amandava amandava in Spain, and the Ruddy duck, Oxyura jamaicensis, from North America. It can be argued that the last example should be eliminated because of risks of hybridization with the native endangered White-headed duck Oxyura leucocephala. Among mammals, some species have been successfully introduced among which the Coypu Myocastor coypus and the Musk rat Ondatra zibethicus may cause damage in wetlands. Attempts to introduce the Floridan rabbit Sylvilagus floridanus as a game species have not been very successful except in some scattered localities. With only 250 naturalized plants in ‘natural’ areas, i.e. 1 per cent of the total plant richness (Quézel et al. 1990), the Mediterranean region is generally considered less vulnerable to invasion by alien plant species

than the other four mediterranean-type regions of the world, but the problem may be slightly underestimated. It is also noteworthy that the flora of the Mediterranean basin often provides an important source of alien species to other mediterranean-type regions, particularly California. The global threats induced by alien plants to rare or endemic Mediterranean plants are still relatively small (Thompson 2005). Indeed, the well-documented survey performed within the project Atlas y libro rojo de la flora vascular amenazada de España (Bañares et al. 2003) mentions that only 8 per cent of the studied populations (n = 2,223) of rare plants are threatened by the competition induced by xenophytes. However, some exotic plant invaders are a serious threat to coastal and riparian ecosystems, with Mediterranean islands being more threatened by invading plants than most mainland areas (Lloret et al. 2004). For example, exotic plant species represent 17 per cent (473 taxa) of the Corsican flora (Natali and Jeanmonod 1996), but only 1.4 per cent (38 taxa) of them are considered as naturalized in the natural vegetation. Nevertheless, numerous alien plants are becoming invasive in some parts of the Mediterranean basin, inducing profound ecological disruptions in natural or semi-natural habitats (Table 23.1). Some South African Carpobrotus taxa (Aizoaceae) were largely introduced at the beginning of the nineteenth century, and are often naturalized, causing one of the most severe invasive threats in Mediterranean coastal and island habitats (Suehs et al. 2001). Studies of genetic variation suggest that C. acinaciformis individuals are of hybrid origin in some islands off the coast of France, and have generated recent introgressed types through repeated backcrossing with C. edulis genotypes (Figure 23.11). All these strains together form a large and variable hybrid swarm with an aggressive invasiveness (Suehs et al. 2004a, b). Thus, the rapid evolutionary potential of hybrid genotypes can constitute a serious additional threat by making hybrid plants particularly invasive. Wetlands are particularly sensitive to invasion and have been colonized by several xenophytes some of which cover huge areas, causing much damage to freshwater plant and animal communities. Examples are Baccharis halimifolia, Cortaderia selloana, Ludwigia grandiflora, and L. peploides, which can form thick carpets completely covering rivers and canals. Coming across an angler recently near a canal in the Camargue, one of us saw the man pulling a bass out of a thick carpet of Ludwigia. It appeared that the belly of the fish was full of Louisiania crayfish. The three players of this scene were three exotic species which presumably developed at the expense of many native species.

Biodiversity and Conservation

629

TABLE 23.1. Some of the most invasive alien plants occurring in the Mediterranean basin Taxa (and main species) Acacia spp. (A. dealbata, A. karroo, A. longifolia, A. melanoxylon, A. saligna . . . ) Agave americana Ailanthus altissima Araujia sericifera Arundo donax Aster squamatus Azolla filiculoides Baccharis halimifolia Buddleja davidii Carpobrotus spp. (C. edulis, C. acinaciformis) Conyza spp. (C. bonariensis, C. canadensis, C. sumatrensis) Cortaderia selloana Cotula coronopifolia Datura stramonium Eucalyptus spp. (E. camaldulensis, E. globulus . . . ) Gomphocarpus fruticosus Ipomoea ssp. (I. sagittata, I stolonifera, I. purpurea, I. indica) Isatis tinctoria Lemna minuta Lonicera japonica Ludwigia spp. (L. grandiflora, L. peploides) Medicago arborea Nicotiana glauca Oenothera spp. (O. biennis, O. glazioviana, O. dillenii) Opuntia ssp. (O. ficus-indica, O. dillenii, O. stricta, O. subulata, O. crassa . . . ) Oxalis pes-caprae Paspalum spp. (P. dilatatum, P. paspalodes, P. vaginatum) Pennisetum villosum Phytolacca americana Pittosporum tobira Ricinus communis Robinia pseudoacacia Senecio angulatus Senecio inaequidens Sorghum halepense Solanum spp. (S. chenopodioides, S. bonariense, S. sisymbrifolium . . . ) Tradescantia fluminensis Xanthium spp. (X. italicum, X. spinosum)

Family

Native distribution

Main invaded habitats in the Mediterranean basin

Mimosaceae

Australia, S. Africa

Woodlands, matorrals, coastal habitats

Agavaceae Simaroubaceae Asclepiadaceae Poaceae Asteraceae Azollaceae Asteraceae Buddlejaceae Aizoaceae Asteraceae

C. America China S. America C. Asia S. and C. America S. and N. America N. America China S. Africa S. and N. America

Coastal habitats, matorrals Riparian or ruderal habitats, woodlands Matorrals, evergreen forests Ruderal and riparian habitats, agroecoystems, old-fields Ruderal habitats, wetlands, old-fields, sandy beaches Freshwater Wetlands, riparian habitats Riparian or ruderal habitats Coastal habitats Ruderal habitats, agroecoystems, old-fields

Poaceae Asteraceae Solanaceae Myrtaceae

S. America S. Africa America Australia

Ruderal habitats, wetlands, old-fields, matorrals Wetlands Ruderal and riparian habitats Woodlands, matorrals

Asclepiadaceae Convolvulaceae

S. Africa S. America

Riparian or ruderal habitats Coastal matorrals, wetlands

Brassicaceae Lemnaceae Caprifoliaceae Onagraceae Fabaceae Solanaceae Onagraceae

S.E. Asia N. and S. America E. Asia S. and N. America E. Mediterranean S. America S. America

Ruderal habitats, old-fields, agroecosystems, matorrals Freshwater Wetlands, riparian habitats Freshwater Coastal habitats, matorrals Coastal and ruderal habitats Ruderal or riparian or habitats, matorrals, old-fields

Cactaceae

C. and S. America

Rocks, matorrals, coastal habitats

Oxalidaceae Poaceae

S. Africa S. America

Agroecosystems, old-fields Agroecosystems, old-fields, wet pastures, wetlands

Poaceae Phytolaccaceae Pittosporaceae Euphorbiaceae Fabaceae Asteraceae Asteraceae Poaceae Solanaceae

N.E. Africa N. America E. Asia Palaeotropical N. America S. Africa S. Africa E. Mediterranean C. and S. America

Ruderal habitats, old-fields Riparian and ruderal habitats, mesophilous matorrals Coastal habitats, matorrals Ruderal or riparian habitats Woodlands, riparian habitats, matorrals Coastal matorrals, evergreen forests Ruderal habitats, agroecoystems, old-fields Ruderal habitats, agroecoystems, old-fields Ruderal or riparian habitats

Commelinaceae Asteraceae

S. America America

Riparian habitats, wetlands, rocks Ruderal or riparian habitats, sandy beaches

Source: Modified from Médail, unpublished data.

The Decline of Mediterranean Biodiversity Except for large mammals and some endemic species on Mediterranean islands, relatively few postglacial extinction events due to human impact have been reported in the Mediterranean in recent times (Chapter 5). This does not mean, however, that the status of Mediterranean

biodiversity is safe. Changes of anthropogenic origin have had many impacts on the distribution and abundance of populations, making a large number decline while others are increasing. In fact, for many plants, some groups of vertebrates, and most groups of invertebrates, micro-organisms and fungi, there is much uncertainty on their status and several supposedly extinct species may survive in unnoticed localities, or as

630

Jacques Blondel and Frédéric Médail

Fig. 23.11. Carpobrotus acinaciformis is a very aggressive invasive plant species in coastal areas (photo: Frédéric Médail).

dormant seeds or bulbs in plants as reported for example by Greuter (1994) who mentioned that several plant species thought to have vanished have recently been rediscovered.

Plant Biodiversity Despite a large number of endemic plant species that are narrowly distributed in a single or few localities, few taxa seem to have become extinct in the Mediterranean region, except perhaps on some islands such as the Canary islands and Madeira, where 24 and 20 per cent respectively of native plant species are considered to be threatened. The first census performed by Greuter (1994) indicated than only thirty-seven Mediterranean plants were presumed totally extinct, i.e. an extinction rate of 0.13 per cent of the native Mediterranean flora (Table 23.2). The present survey, taking into account several recent floristic checklists and red data books, allowed us to identify forty-two species and subspecies that are probably extinct within the strict Mediterranean bioclimatic region, which represents an

extinction rate of 0.15 per cent. Sixteen of these taxa were not cited by Greuter (1994) and only half of them are common in the two surveys. The largest number of plant extinctions have been reported for Turkey (10 taxa), Greece, and Italy (6 taxa each). However, the number of documented cases of extinctions in countries of North Africa and the Middle East is presumably disproportionately low given the ongoing huge loss of habitats. The current level of extinction is perhaps not significantly higher than during earlier geological periods (Chapter 4), but it is very difficult to estimate the magnitude of postglacial extinctions for plants. Among the very few cases of local extinctions during historical time, the disappearance of Tetraclinis articulata stands from Cyrenaica (Libya) was already testified by Theophraste in c.500 BC. This tree of the Cupressaceae family occurs today only in the Maghreb—its native status is doubtful in Andalucia. The strong local persistence and resilience of Mediterranean plant species is mainly due to their life-strategies, with a high tolerance of stress and disturbance (Chapter 7). One example is the

Biodiversity and Conservation

631

TABLE 23.2. Threatened vascular plants by country based on the former IUCN categories and included in the 1997 IUCN Red List of Threatened Plants Country

Ext

Ex/E

E

Albania Algeria Croatia Cyprus Egypt France Gibraltar Greece Israel Italy Jordan Lebanon Libya Malta Morocco Portugal Spain Syria Tunisia Turkey

1

3

1 31

2 7 6

3 1 2

1 1 1 1 2 3 2 10

3

1

9 19 21 28 7 29 1 2 4 3 46 185 5 1 47

V

22 2 13 8 81 3 80 7 80 1 18 1 3 113 272

167

R

I

73 80 4 23 45 83 1 430 13 190 6 1 30 10 157 98 484 3 18 161

2 8

Total threatened plants

Total plant species

79 141 6 51 82 195 4 571 32 311 9 5 57 15 186 269 985 8 24 1, 876

6 10 7 32 3 12 1 2 5 4 23 12 41 5 53

% threatened

3, 031 3, 164 5, 347 1, 682 2, 076 4, 630 — 4, 992 2, 317 5, 599 2, 100 3, 000 1, 825 914 3, 675 5, 050 5, 050 3, 000 2, 196 8, 650

2.6 4.5 0.1 3 3.9 4.2 — 11.4 1.4 5.6 0.4 0.2 3.1 1.6 5.1 5.3 19.5 0.3 1.1 21.7

Note: The figures also include the plant species that are present in the non-Mediterranean areas of these countries. Source: Modified from Walter and Gillett 1998.

amazing persistence of only c.200 individuals of Zelkova sicula, a recently discovered Tertiary relict tree located in only one small area in south-east Sicily (Garfi 1997) (Figure 23.12). This extreme case of persistence through time of a plant species results from its vegetative propagation, which combines a clonal structure and a process of adaptive reiteration with a continuous replacement of the vegetative apex which is repeatedly destroyed by drought or grazing. Even if it is very difficult to have a precise idea of the number of plants that are threatened in the Mediterranean, some details are available for certain areas or plant categories. In the southern shore countries, which are still in a process of ongoing degradation as explained above, Leon et al. (1985) estimated that nearly 25 per cent of the Mediterranean flora may be threatened in the decades to come, but precise and updated appraisals are still lacking. The unique flora of Mediterranean islands is on the whole threatened, especially the endemic species that grow on coastal and low-altitude habitats: on large islands, the percentage of taxa that are threatened on a global scale ranges from 2 per cent (Corsica) to 11 per cent (Crete) (Table 23.3). A detailed comparison of life-form spectra for the floras of south-east France and Corsica in relation to altitudinal distribution and rarity (Verlaque et al. 2001), indicates that endemic plants

are not more prone to extinction than other species. Extinction rates and rarity percentages of endemics in Corsica are lower than those of non-endemic taxa (1.6 v. 3.8% and 33 v. 37.5% respectively). Rare plants (i.e. occurring in less than ten localities) are mainly located at low altitude, a sensitive zone with the highest extinction rates. As much as 87 per cent of the extinct plants in Corsica and 90 per cent in south-east France occurred between 0 and 800 m above sea level, mainly in arable fields, wetlands, coastal areas, and rocky grasslands.

TABLE 23.3. Threatened vascular plants on the seven large Mediterranean islands, based on the former IUCN categories Island

Balearic Corsica Sardinia Sicily Crete Malta Cyprus

Ext

E

V

R

I

Total

Threatened taxa/total flora (%)

1 1 0 1 0 1 0

10 8 11 11 11 0 9

14 27 30 26 61 1 14

43 10 21 45 118 10 22

1 1 1 4 3 4 6

69 47 63 87 193 16 51

5 2 3 3 11 2 3

Note: Ext: extinct; E: endangered; V: vulnerable; R: rare; I: insufficiently documented. Source: Delanoë et al. 1996.

632

Jacques Blondel and Frédéric Médail

Fig. 23.12. Ramet of Zelkova sicula (Ulmaceae), a relict and very threatened palaeoendemic small tree located in south-eastern Sicily (from Raimondo et al. 1994 in Quézel and Médail 2003).

The global extinction rate of plant species is higher in south-east France (4.79%, n = 161 local extinct plants) than in the less disturbed island of Corsica (3.48%, n = 74 local extinct plants). This is probably due to the more ancient and important human impact on continental areas, which include 7.3 times more inhabitants. On the whole, plants without visible vegetative organs during unfavourable seasons (annuals, aquatic, and bulbous groups) are the most threatened, with a rate of population extinction reaching 21 per cent for aquatic plants in south-east France (ibid.). Considering Mediterranean trees, the World List of Threatened Trees (Oldfield et al. 1998) includes thirtyfive Mediterranean species, and the 1997 IUCN Red Data Book mentions twenty-nine tree species (Walter and Gillett 1998), but a more precise list includes sixty-one taxa with forty-two endemics (Quézel and Médail 2003),

some of them being progenitors of cultivated trees (Malus, Olea, Phoenix, Prunus, Pyrus). One of the most noteworthy of these is Phoenix theophrasti, an eastern Mediterranean palm recognized by Theophrastus and Pliny but only formally described in 1967 by Greuter. According to ancient texts, the former distribution of this palm was more extensive, but nowadays this relict tree persists only in scattered populations at low altitude (0–350 m a.s.l.), in riparian habitats, and moist cliffs within some coastal sites in Crete, the Peloponnese, and south-west Turkey (Boydak 1987). Several populations are threatened by the continuous increase of tourist pressure along the coasts of Turkey and Greece, especially water harnessing, and by the putative increase of drought events associated with global warming (Chapter 3). Threats upon the emblematic cedar species are also evident (Quézel and Médail 2003). The famous Lebanon cedars were exploited for several millennia by the Egyptians, Phoenicians, Assyrians, Romans, and Turks who used large volumes of timber for ship-building and other uses. Only fourteen isolated and fragmented stands of Lebanon cedar currently remain, covering at most an extremely reduced surface area of 2,700 ha. The conservation of bulbous plants is another concern around the Mediterranean. In the Near East, notably in Turkey, intensive harvesting of geophytes for horticultural or culinary purposes still constitutes a serious problem since many of these taxa are endemic. In the 1980s in Turkey alone, 60 to 80 million bulbs (mainly Galanthus, Eranthis, Leucojum, Anemone, and Cyclamen) were collected in the wild each year, and 57 million tubers belonging to thirty-eight species of orchids were picked annually for the preparation of salep, a popular milk drink (Sezik 1989). Some integrated conservation programmes including in situ bulb propagation and then resale by villagers have reduced by half these huge removals and constitute an alternative source of bulbs for the international trade (Entwistle et al. 2002). The narrow distribution of many Mediterranean geophytes increases their extinction risks, even if storage organs such as bulbs or rhizomes may enhance population persistence over time. For example, the western Mediterranean snowflakes (Acis, Amaryllidaceae) encompass nine bulbous plant species and almost all of these are endemics that are threatened by various factors, including severe human impacts such as overgrazing and deforestation. Acis tingitana, which is a north Moroccan endemic, is especially vulnerable to such impacts. Other kinds of land use change that encourage rapid habitat colonization by competitive herbs and shrubs are serious threats for Acis nicaeensis and Acis fabrei (Figure 23.13),

Biodiversity and Conservation

633

agricultural landscapes, and the putative extinction of several endemic or relict plants, with several rare native cytotypes. On the whole, however, data allowing a predictive overview of plant extinctions in relation to patterns of rarity and threats in the Mediterranean are still lacking. Future work should be inspired by the detailed study on Centaurea corymbosa, a narrow endemic and cliff-dwelling Asteraceae from southern France (Colas et al. 1997, 2001). The low seed dispersal and reduced ability to establish new populations due to obligatory cross-pollination (allogamy) and a single reproductive event in the life of the plant (semelparity) represent a serious handicap, enlarged by an Allee effect (negative effect of low population density on demographic parameters) on seed production. As the metapopulation dynamics of this species is very slow and depends upon only six natural populations located within an area of 3 km2 , the survival of this species is indeed in the balance.

Animal Biodiversity

Fig. 23.13. Acis fabrei, a narrow endemic with only four known populations from the southern slopes of the Mont Ventoux (Vaucluse, south-eastern France) (photo: Frédéric Médail).

which are narrow endemics of southern France, and Acis longifolia, which is an endemic from northern Corsica. Mediterranean segetals are plant species, mostly annuals, that are associated with cereal crop fields. They encompass over 1,500 species but their decline is also worrying as a result of land use changes that have induced profound shifts in the floristic composition of arable fields, olive groves, and other traditional agricultural lands. Changes to cultural practices can favour the spread of invasive weeds, mainly polyploid plants characterized by a high competitive capacity, at the expense of the traditional segetals. The latter are often stress-tolerant and annual species, autogamous, diploid, and with an autumnal germination (e.g. Camelina, Cephalaria, Delphinium, Garidella, Hypecoum, Nigella, Tulipa). The disappearance of this highly specialized flora contributes to the decline of Mediterranean

Historical records, both palaeontological and archaeological, show that a large human-induced decline in animal biodiversity started many thousands of years ago in most parts of the Mediterranean. The end of glacial times was characterized in the Northern Hemisphere by the presence of an extraordinary large number of large mammals including no less than twenty-five herbivorous as well as predatory species (Table 23.4). Most of them are now extinct but a testimony of this magnificent Late Pleistocene mammal fauna is provided by dozens of fossil sites in southern Europe as well as by the superb wall paintings in many ornate caves of southern France and northern Spain. Two caves with beautiful paintings have recently been discovered in southern France, the Chauvet cave in 1994 and the Cosquer cave in 1996. The Chauvet cave contains some of the oldest (about 30,000 BP) and most beautiful cave paintings ever found in the Mediterranean basin (Chauvet et al. 1995). A vast bestiary is portrayed, giving a vivid picture of what the fauna of this epoch looked like, including three hundred or more different animals such as bisons, horses, bears, deer, mammoths, hyenas, panthers, lions, rhinos, reindeers, aurochs, and ibexes, as well as the only known representation in Palaeolithic art of large birds such as the Eagle owl. In the underwater Cosquer cave near Marseilles, paintings depict many mammals such as horses, bisons, and ibexes, and an extraordinarily detailed rendition of the Great auk. This painting, which dates back 20,000 years, testifies that breeding colonies of this species, which is now extinct, occurred on the

634

Jacques Blondel and Frédéric Médail TABLE 23.4. List of large mammals that were present in the Mediterranean basin during the Late Pleistocene, including species found as fossils in various deposits of southern France and which became extinct (E) before the Holocene Families and species Canidae Canis lupus Vulpes vulpes Alopex lagopus Cuon alpinus Ursidae Ursus thibetanus Ursus arctos Ursus spelaeus Hyaenidae Crocuta spelaea Felidae Panthera (Leo) spelaea Panthera pardus Lynx spelaea Felis silvestris Mustelidae Gulo spelaeus Meles meles Proboscidae Palaeoloxodon antiquus Mammuthus primigenius Rhinocerotidae Coelodonta antiquitatis Dicerorhinus hemitoechus

Late Pleistocene

Extinct

+ + + +

+ + E E

+ + +

E + E

+

E

+ + + +

E + E +

+ +

E +

+ +

E E

+ +

E E

Families and species Equidae Equus sp Equus germanicus Suidae Sus scrofa Cervidae Rangifer tarandus Megaceros giganteus Cervus elaphus Capreolus capreolus Bovidae Bos primigenius Bison priscus Hemitragus sp. Hemitragus cedrensis Capra sp. Capra ibex Rupicapra rupicapra Rupicapra sp. Sciuridae Marmotta marmotta Castor fiber

Late Pleistocene

Extinct

+ +

+ E

+

+

+ + + +

E E + +

+ + + + +

E E E E +

+ +

+ E

+ +

+ +

Source: Modified from Defleur et al.1994 in Blondel and Aronson,1999.

shores of the Mediterranean Sea during glacial times. Some survivors of this ancient fauna, which included a surprisingly high number of large predators, disappeared only recently, since the lion and elephant survived until historic times in Greece and Syria, respectively. Perhaps the most dramatic and still controversial event of the Late Pleistocene is the tempo and mode of the mass extinction of a large part of the rich megafauna that was so characteristic of this epoch (Chapter 5). The question whether this mass extinction was mainly caused by humans as advocated by the overkill hypothesis (Martin 1984) or resulted from dramatic environmental changes, or a combination of both, is still open. As pointed out by Martin and Klein (1984), large body size repeatedly appears as a characteristic associated with higher rates of extinction during episodes of environmental change. Extinction events in Europe and North America during the Late Pleistocene differentially eliminated mammals of large body size and open steppe and grassland. Birds that disappeared at the same time also included a disproportionate number of carnivores and scavengers as well as species of large size and open habitats. The decimation of large endemic mammals continued well into the Holocene, however, as sadly

illustrated by the human-induced extinction of all the large ‘mega-nano-mammals’ of Mediterranean islands, for example the dwarf hippos and elephants of Cyprus, Malta, Sicily, following the colonization of these islands by humans in the early Holocene, some 10,000 years ago (Simmons 1988, 1991). In addition to these mammals, tortoises and flightless owls which populated most Mediterranean islands were also decimated by humans as soon as they invaded the islands (Blondel and Vigne 1993; Vigne et al. 1997). In more recent times, the combination of habitat changes and direct persecution has doomed to extinction many species of large mammals, especially in North Africa. For example, the Brown bear Ursus arctos, the Ass Equus asinus, several species of gazelles (Gazella rufina, Oryx dammah, Alsephalus busephalus), the Lion Panthera leo, the Porcupine Hystrix cristata, the Cuvier’s gazelle Gazella gazella, and the Mouflon Ammotragus lervia have been extirpated from many parts of the Mediterranean, especially North Africa (see Blondel and Aronson 1999). Species that suffer directly from persecution are top predators such as the Panther Panthera pardus and the Monk seal Monachus monachus which are on the verge of extinction. The wolf, which is the

Biodiversity and Conservation

635

Fig. 23.14. Three taxa of salmonids: (top) the Marble trout which is the most endangered fish species of the Adriatic basin, (middle) the Corsican trout with pure genotypes only in the upper reaches of river basins, and (bottom) the Brown trout which is still widespread across the region.

animal most closely associated with wilderness areas, has already disappeared from fourteen European countries and is in danger of disappearing from several others. Viable populations still persist in the Balkans and, with 500–700 individuals, Greece has the second largest population in the European Union after Spain. The species is currently reoccurring in France following immigration from Italy in the 1990s. Freshwater fish are perhaps the most interesting group of Mediterranean vertebrates, but also one of the less known and the most threatened with more than 300 species, 44 per cent of them being endemic (Crivelli and Maitland 1995). Several families include a large number of endemic species, for example the eighty-

three species of Cyprinidae, twelve species of Cobitidae, and eight species of Salmonidae (Crivelli 1996), with beautiful but endangered species such as the Marble trout Salmo marmoratus and the Corsican trout Salmo trutta macrostygma (Figure 23.14). Unfortunately, 70 per cent of these endemic species are seriously threatened, and four species are already extinct (ibid.). Most threats are from the degradation of water both in quality (pollution, eutrophication) and quantity (huge amounts of water are pumped for domestic use and irrigation), hybridization and invading species, especially predatory species. Other threats to freshwater animals are habitat destruction and loss combined with pollution, which are particularly harmful to the troglobiotic (cave-dwelling)

636

Jacques Blondel and Frédéric Médail

fauna along Dalmatian coasts in countries bordering the Adriatic Sea, especially in Slovenia and Croatia. In limestone areas, cave waters as well as interstitial waters are contaminated by diffuse pollution from the use of fertilizers, pesticides, and, episodically, spills (Sket 1999). Commercial collecting of cave beetles and of one of the most fascinating Mediterranean amphibian, the blind cave salamander Proteus anguinus, can also be locally important threats. Annexes of the EU Habitat Directive listing species of ‘European Importance’ tend to ignore this kind of fauna, which is particularly diverse and specific to the karst areas of Mediterranean Europe (Chapter 10). Reptiles and amphibians have not suffered serious extinction events in recent times, since only the Crocodile Crocodilus niloticus and the endemic Blackbelly toad Discoglossus nigriventer have disappeared from Israel and southern Morocco, and marshy areas of the Israel–Syria border, respectively (Blondel and Aronson 1999). However, most populations of amphibians and many populations of reptiles are declining

everywhere in the world, including the Mediterranean (Balmford et al. 2003; Grillas et al. 2004). The main threats include epidemics, habitat acidification, desiccation, climate change, and the introduction of exotic species. For example, the massive introduction of Slider turtles Trachemys scripta elegans into Europe as pets often induces the release of these exotic turtles into natural habitats where they have been successful in reproducing. Competition with the European pond turtle Emys orbicularis results in weight loss and high mortality in the latter where the two species co-occur (Cadi and Joly 2004; Figure 23.15). Another threat for amphibians is the introduction of predatory fish species in amphibian breeding sites which are naturally fish-free. In such sites amphibians develop inducible defences that are costly to produce, potentially affecting the population dynamics of species (Teplitsky et al. 2003). Among birds, no more than three large species in North Africa (the Lappet-faced vulture Torgos tracheliotus, the Helmet guinea fowl Numida meleagris, and the Arabian bustard Choriotis arabs) have disappeared, while

Fig. 23.15. The pond terrapin Emys orbicularis, a vulnerable species in decline (photo: H. Hafner).

Biodiversity and Conservation

637

Fig. 23.16. A pair of Bonelli’s eagles Hierraaetus fasciatus at their nest. This species is a typical cliff-dwelling raptor which is declining in most parts of its range (photo: Jacques Blondel).

a few others such as the Demoiselle crane Anthropoides virgo and the Bald ibis Geronticus eremita are in danger of extinction. There are, however, many changes in the structure and composition of the bird fauna of the Mediterranean basin. Many species that were formerly widespread everywhere in the basin are now confined to small, localized populations, many of them being threatened, mostly as a consequence of habitat loss or change, pollution, and sometimes food shortage. A serious decline in bird populations has been reported for many species and this decline may be catastrophic for some groups such as farmland birds (Donald et al. 2001). Some species are particularly vulnerable, for example the Bonelli’s eagle Hierraaetus fasciatus (Figure 23.16) and the Lesser kestrel Falco naumanni with a decline that has accounted for 95 per cent of the European population since 1950 (BirdLife International 2000). For large insectivorous birds such as the Lesser kestrel and several species of shrikes, the main cause of decline is food shortage resulting from the intensification of agriculture and urban sprawl (Liven-Schulman et al. 2004). For migrant birds, pesticide use in Africa to fight against

locusts might affect birds through a sharp reduction in prey availability. On the whole, there is a general trend of decline in population sizes, especially waders, sub-Saharan migrants, and farmland species. Despite some progress made in the European Union through the Birds Directive, the conservation status of many birds in Europe has worsened since 1994, and 43 per cent of Europe’s 526 species are now under threat (BirdLife International 2004), which is 5 per cent more than in 1994 (Tucker and Heath 1994). Socio-economic and political changes may also seriously affect some components of biodiversity as illustrated by the story of Griffon vultures and transhumance. The Griffon vulture Gyps fulvus used to belong to a kind of trophic community closely linked to migrating herds of wild mega-herbivores such as the European Bison, Auroch, Mouflon, Wild goats, Saiga antelope, and Tarpan. The up and down transhumance of these animals between lowlands in winter and highlands in summer occurred annually in April–May and October. The carcasses and dung of these animals— which were scattered along migration routes—provided

638

Jacques Blondel and Frédéric Médail

plenty of food for vultures, corvids, mammals, insects (e.g. dung beetles), and other animals associated with these resources. Transhumance of domestic herds of sheep and cattle probably followed the same transhumance routes as wild herds and persisted in Mediterranean Europe up to the middle of the twentieth century, providing plenty of food for vultures and associated animals in many parts of the Mediterranean. For example, at least nine large colonies of the Griffon vulture have been reported to occur along the gorges of the Drina River in the Balkan Peninsula, as shown by many sub-fossil remains spreading from the upper Pleistocene and Holocene to modern times (Marinkovic and Karadzic, 1996). Political circumstances with changes in borders combined with shorter migration routes and sedentarization of cattle herds resulted in a decline and abandonment of transhumance. In the 1960s, laws forbidding transhumance in the former communist federal state of Yugoslavia further dramatically reduced the abundance of nomadic herds which subsequently completely disappeared. The break-up of the long migratory routes of cattle on the Balkan Peninsula started at the end of the nineteenth century after the collapse of the Ottoman Empire and the formation of new national borders. This, combined with a widespread use of poison (e.g. strychnine) against large carnivorous mammals, resulted in a complete collapse of the animal communities associated with these long-distance mammal migrants. On the whole, for birds as well as for many insects and reptiles, the main consequences of human-induced changes in Mediterranean habitats over time have not been so much a decrease in overall species richness at a regional scale than a tremendous advantage for species adapted to drylands and shrublands, as opposed to forest-dwellers. In some cases, direct persecution resulted in the death of millions of birds. Particularly sad examples are the large-scale destruction of migrant birds in Mediterranean islands such as Malta, Cyprus, and others where migrating birds, including the tiny kinglets, used to provide a providential amount of food for local people twice a year (Fenech 1993). Magnin (1991) has estimated that up to 1,000 million birds are killed annually in the basin when they stop over during migration. Except for very few exceptions such as butterflies and some large families or genera of conspicuous beetles, almost nothing is known on the distribution and abundance status of most groups of invertebrates. For groups that have been studied in some detail, there is a growing body of evidence for a severe decline (Collins and Thomas 1991; Pullin 1995). For example, a dramatic

decline has been reported for large conspicuous insects such as butterflies, large bees, dragonflies, and many groups of beetles, including all saproxylophagic groups that are narrowly tied to decaying wood (Speight 1989). The decline in population abundance of most groups of large insects paralleled that of large insectivorous birds such as the Roller Coracias garrulus, the Little owl Athene noctua, the Scops owl Otus scops (Figure 23.17), and all species of shrikes. Another dramatic decline of invertebrates is that of earthworms (Granval and Muys 1992) which has been attributed to leaching, acid rain, heavy metal pollution, and all factors associated with intensive mechanized agriculture. Given their key function in ecosystems, the decline of earthworms and the concomitant soil compaction contributes to an increase in soil erosion and severe flooding (Chapter 6).

Genetic Changes and the Decline of Human-selected Plant and Animal Varieties Subtle changes potentially detrimental to biodiversity may result from (1) introgression between native and cultivated populations as is the case with some northern populations of Oleander Nerium oleander, or (2) hybridization between closely related species that secondarily come in contact as a result of human-induced changes in the environment. For example, Mediterranean orchids (Ophrys, Orchis, Dactylorhiza, and Serapias) have a wide range of isolating barriers that lower the risks of interspecific recombination under normal circumstances. But in many regions of the basin, isolating barriers may have become ineffective because of millennia of human activities that have altered landscape structure and created disturbed sites conducive to hybridization. On the other hand, a recent aspect of biodiversity decline is that of the regression of species, cultivars, and various breeds of economic value that are now neglected after centuries of selection by humans. The Mediterranean basin is a key area for the presence of many wild progenitors of cultivated plants and for the origin and spread of agriculture (Zohary and Hopf 2000). In many species, adaptive intraspecific variation occurred as a response to human-induced selection and habitat changes over millennia, resulting in the differentiation of a burst of local ecotypes and gene pools in plant species, with region-specific characters fitting them to local climate and environmental conditions (Chapter 7). Vavilov (1935) enumerated a large number of plant taxa that differentiated as a result of human selection through

Biodiversity and Conservation

639

Fig. 23.17. The Scops owl Otus scops is threatened by the decline of large invertebrates which constitute the bulk of its food (photo: Jacques Blondel).

the process of domestication. Over the centuries, hundreds of varieties of olive, almond, wheat, grape, etc., were passed down if not preserved in hortus. In many cases, the horticultural and agricultural selections and discoveries of past generations have been lost in recent decades. For example, only a very few of the 382 named cultivars of almond that were in use at the turn of the last century on the island of Mallorca are left (Socias y Company 1990). Throughout the Mediterranean area, the great majority of ‘minor’ fruits, nuts, vegetables, and other plant varieties selected in the past are extinct and

lost forever. The loss of these ancient varieties is an issue of great concern and much effort is devoted today to restore and preserve them from extinction. In the same way, another threat that directly concerns humans is the extinction of many livestock breeds or varieties. Of 3,381 breeds or varieties (horses, donkeys, cows, buffalo, pigs, goats, and sheep) recorded worldwide up to the twentieth century, 618 (16%) were already extinct—many of them in the Mediterranean, which is the region with the greatest variety of breeds. The Mediterranean basin includes 45 per

640

Jacques Blondel and Frédéric Médail

cent of the bovid varieties and 55 per cent of the goat varieties of Europe and the Middle East but a large proportion of them are in danger of extinction (Georgoudis 1995), mainly because European dairy breeds dramatically outnumber of local breeds. Greece still has a total of thirty-nine livestock breeds, for example, but has already lost eight in the last century (Catsadorakis, 2003).

Conservation Framework and Programmes for Conserving Biodiversity and Restoring Endangered Species and Populations Many protected areas have been created around the Mediterranean basin, with various levels of protection, and there are currently seventy-six National Parks and forty Man and Biosphere Reserves (Table 23.5). Much effort has also been done in recent years to protect wetlands in the framework of the Ramsar convention, and there are at present 206 Ramsar sites representing a total area of more than 4 million hectares (Table 23.6; Chapter 9). Protected forest areas amount to c.10 per cent of forest coverage in the north shore countries of the MediterTABLE 23.5. Major protected areas such as National Parks (n = 76) and Biosphere Reserves (n = 40) within the Mediterranean bioclimatic region Country

National Parks (no.)

Albania Algeria Croatia Cyprus Egypt France Greece Israel Italy Lebanon Libya Morocco Portugal Spain Tunisia Turkey

3 6 5 1 0 2 10 3 15 1 3 7 0 5 6 9

Surfaces of National Parks (ha) 6,260 75,867 51,964 9,337 0 71,975 68,732 12,504 986,504 810 150,000 293,000 0 187,856 40,026 202,953

Biosphere Reserves (no.) 0 5 1 0 1 4 2 1 6 0 0 2 1 13 4 0

Area of Biosphere Reserves (ha) 0 152,806 200,000 0 7,000 296,585 8,850 26,600 256,135 0 0 975,4151 358 1,334,334 73,562 0

Source: Completed after Ramade (1997), Quézel and Médail (2003) and <www. unesco.org/mab/>, accessed 17 November 2008

TABLE 23.6. List of Ramsar sites within the Mediterranean bioclimatic region Country

Ramsar sites (no.)

Albania Algeria Croatia Cyprus Egypt France Greece Israel Italy Lebanon Libya Malta Monaco Morocco Portugal Slovenia Spain Syria Tunisia Turkey

3 39 1 1 2 5 10 2 42 4 2 2 1 24 17 2 36 1 1 11

83,062 2,897,115 11,500 1,585 105,700 136,552 163,501 366 55,489 1,075 83 16 10 272,010 73,784 955 156,351 10,000 12,600 157,782

206

4, 139, 536

TOTAL

Total area of Ramsar sites (ha)

Source: Modified from <www.ramsar.org>, accessed 1 April 2007 (see Chapter 9).

ranean but hardly more than 3 per cent is effectively protected, at least on paper. In fact most protected forested areas are too small and isolated to provide real security for habitats and their associated plant and animal communities. In some regions, for example in Greece, Croatia, Bosnia-Herzegovina, and Slovenia, almost 50 per cent of the land is still forested, sometimes with beautiful, nearly pristine forests as in parts of the Peloponnese and the Rhodope mountains in Greece. The Vikos-Aoos National Park in Epirus, northwest Greece (Figure 23.18), is still in a beautiful state of wilderness with Lynx, Bear, Wolf, Otter, Chamois, Roe deer, Wild boar, and Wild cat (Catsadorakis 2003). In contrast, no more than 1 per cent of French forests can be considered in a satisfactory state of conservation (Vallauri 2003). The situation in North Africa is even worse with an ongoing degradation and reduction of the forest cover without any effective conservation policy (Chapter 20). In contrast to vertebrates, which have enjoyed a long tradition of conservation, the protection and restoration of threatened plant populations is less advanced in the Mediterranean region, and many reintroduction programmes have unfortunately failed. This is probably due to a lack of preliminary integrative biological

Biodiversity and Conservation

641

Fig. 23.18. The Vikos Gorge with its spectacular limestone cliffs and, in the middle distance, forested scree slopes and valley floor in the Vikos-Aoos National Park in Epirus, north-west Greece (photo: Jamie Woodward).

and ecological studies both on the target plants and their potential habitats. One interesting initiative for providing protection to threatened flora and endemic plant species (e.g. Antirrhinum pertegasii, Silene diclinis, Verbascum fontqueri) is the Plant Micro-Reserves (PMR) network of the Valencian region launched in 1994 by the Generalitat Valenciana (Laguna et al. 2004). This network of small (2–20 ha) reserves currently includes 207 PMRs which confer protection to 57 per cent of the critically endangered (CR sensu IUCN) plants present in this region. In France, several reintroduction programmes are carried out by official structures such as Conservatoires Botaniques Nationaux with some successful outcomes, notably in Corsica. Significant results have also been obtained within the framework of the European Union LIFE programmes, for example concerning the practical tools to restore populations of

a very rare west Mediterranean amphibious quillwort (Isoetes setacea) in temporary flooded pools threatened by the colonization of several shrub and tree species (Rhazi et al. 2004). A crucial point to ensure a rational conservation network is to identify biogeographic areas with both representative and high biodiversity value. Following the Important Bird Areas (IBAs) model developed by BirdLife International and covering 3,619 sites in fifty-one European countries, Plantlife International and IUCN (2003) have recently launched the Important Plant Areas (IPAs) programme in the Mediterranean region. The aim of this initiative is to identify areas that are appropriate for a site-based approach for conservation and to protect a network of the sites that are the most relevant for wild plants throughout the Mediterranean region. This network of IPAs should provide the framework for governments to achieve the Target 5

642

Jacques Blondel and Frédéric Médail

of the Convention on Biological Diversity’s Global Strategy for Plant Conservation, which calls for the protection of 50 per cent of the most important areas for plant diversity. Three main criteria have been used to identify IPAs: (1) the site holds significant populations of species of global or regional concern (presence of threatened species); (2) the site has an exceptionally rich flora in a regional context in relation to its biogeographic zone (species richness); (3) the site is an outstanding example of a habitat type of global or regional concern (presence of threatened habitats). Some Mediterranean IPA projects are already completed, for example in Turkey (2002), and in Croatia (2003) where many of the eighty-eight identified IPAs lie outside the existing protected areas. Vertebrates, especially birds, provide many examples of a steady biodiversity recovery after restoration and proper management of habitats. This is the case in some wetlands in spite of an ongoing loss of marshland in the Mediterranean as a whole. In the Camargue, for example, the diversity and wealth of bird communities has dramatically increased in recent times as a result of an increase in the size of the populations of several species (herons, gulls, terns) and the appearance of new breeding species (e.g. the Spoonbill Platalea leucorodia, the Great white egret Egretta alba, the Glossy ibis Plegadis falcinellus, the Greylag goose Anser anser, the Cormorant Phalacrocorax carbo, and several others). Another spectacular example of biodiversity recovery is that of the Moulouya River mouth in northern Morocco (Brosset 1990). The size of this marshland has increased tenfold thanks to the building of large reservoirs in the upper course of the river and the expansion of irrigated farmland that raised the level of the water table. This water management strategy was initially introduced to improve local agriculture. It also indirectly increased the area of marshland and benefited many species so that the area is today one of the best breeding sites in North Africa for many rare species such as the Purple gallinule Porphyrio porphyrio, the Marbled teal Marmonetta angustirostris, and many others. Encouragingly, this example is a clear demonstration that some aspects of biodiversity can recover quickly as soon as favourable conditions reappear. Hundreds of species could be reinforced or locally reintroduced and many projects have been designed to preserve, reintroduce, or restore populations of particularly endangered species, including plant and animal breeds that had been selected by humans. For example, a project to reintroduce the Griffon vulture Gyps fulvus in the Cévennes, France, succeeded in the reestablishment—one century after its local extinction—

of a self-sustaining population of this species that today includes more than 200 individuals (Terrasse 1996). Other reintroduction projects for this species and for the rarer Black vulture Aegypius monachus elsewhere in southern France promise to be successful. The World Wildlife Fund launched several projects to reintroduce locally extinct populations of several species of vultures, especially the Black vulture Aegypius monachus in Greece (in the Dadia forest) and the Lammergeier Gypaetus barbatus in several countries of Mediterranean Europe. Among the mammals, the Sardous deer, a local population of the Red deer Cervus elaphus, has been successfully saved in Sardinia and recently reintroduced in Corsica. Several ongoing programmes contribute to reinforce endangered species such as the Iberian lynx Lynx pardus in Spain, as well as several species of mammals in Israel, e.g. the Mountain gazelle Gazella gazella. More stringent regulation policies and active action plans launched by BirdLife International for the protection of raptors succeeded in the restoration of many populations of endangered species (Muntaner and Mayol 1996). Reintroduction policies are faced with many problems, however. If properly executed on scientific grounds, they are extremely expensive and require clear strategies to raise public awareness. In addition, economic or political criteria often prevail over scientific arguments for biological conservation because it is much easier to raise funds for protecting flagship species to which public attention can be drawn than for enhancing the diversity of inconspicuous soil micro-invertebrates. In the semi-arid and arid south and east margins of the Mediterranean, successful projects currently include the restoration of the Oryx Oryx dammah and several species of gazelle and deer in several countries (e.g. Tunisia, Morocco, Jordan, Israel, and Syria). There is also currently much effort to preserve the myriad of ecotypes and gene pools of cultivated plants and domesticated animals that have been selected over millennia in all parts of the basin (Diamond 2002). One challenge is to identify and save these wild ancestral gene pools before they are lost for ever. International, governmental, and private organizations currently combine their efforts to preserve these genetic resources of old breeds of cattle, sheep, goats, and the Mediterranean buffalo. In the same way, many programmes aim to conserve genetic resources of fruit trees, grapes, field crops, forage species, vegetables, and ornamental plants (Charrier 1995). In this context, one promising initiative is the International Centre for Advanced Mediterranean Agronomic Studies (CIHEAM) which includes several institutes in Italy (Bari), Greece (Chania), France (Montpellier), and Morocco (Meknès). These Institutes

Biodiversity and Conservation

carry out research involving both in situ and ex situ study of local races of traditional Mediterranean plants. One aim of these institutes is to help promote sustainable development programmes using these local ecotypes.

The Uncertain Consequences of Global Change for Mediterranean Biodiversity Most of the ups and downs in Mediterranean living systems are closely linked with human population pressures which have changed many times through the long common history of ecological systems and human societies. The dynamics of human populations will be a major factor in the future of the Mediterranean. The northern countries which contributed about two-thirds of the human population in the Mediterranean basin in the 1950s will contribute only about one-third in 2025 as a result of demographic growth in the southern and eastern countries of the basin. By 2025 in these countries, from Morocco to Turkey, the human population will have become five times that of 1950 (Grenon and Batisse 1989). In coastal areas which are particularly sensitive to disturbance, the projected changes in human population are 200 million in 2025 instead of 145 million in 2000, with 170 million instead of 95 million living in urban areas. In this context, the future of biological diversity will depend on how humans societies will learn to live together and with their natural heritage in the forthcoming decades. The Mediterranean is the most popular tourist destination in the world, which means that conservation and proper management of coastal areas in particular will be one of the most serious challenges in the forthcoming decades (Chapter 13). This will be especially important on the 18,000 km of coast on the Mediterranean islands, all the more so because in the future, tourism will be the major resource in many countries (Benoit and Comeau 2005). Therefore, in addition to natural resource conservation and management, priority should be given to the protection of coastal areas and wetlands where the strongest pressures occur. The various components of global change, especially climate changes, are other challenges to meet. Scenarios of the IPCC (2001) show that the Mediterranean region will be particularly affected by global warming and a decrease in rainfall. Water shortage in the future could be critical in a region with increasing demand for water supply for an ever-growing population in large cities and the enormous impact of tourism

643

(Chapter 21). In addition, deforestation and increasing rates of runoff due to deforestation and soil erosion in North Africa will accelerate the process of desertification (Chapter 20). With an expected temperature increase of c.3–5◦ C in the Mediterranean during the twentyfirst century, potential evapotranspiration should on average reach 200 mm annually, which is equivalent to a loss of 50 mm in annual rainfall (Le Houérou 1990). The expected shifts in vegetation belts resulting from increased aridity will be an upward move of 80–100 m in altitude and 50–80 km northwards in latitude. As a whole, the geographical distribution of the natural vegetation will undergo a shift towards more thermophilous types. On the other hand, if the predictions are accurate, productivity will increase as a result of the extension of the growing season by one to two weeks (Penuelas et al. 2002), and the reduction in winter cold stress. Furthermore, with a doubling of livestock between 1950 and 1986 and a current density of about one sheep-equivalent per hectare against an overall carrying capacity of one sheep-equivalent per 10–12 ha, there will inevitably be a severe encroachment of desert-like conditions over the entire arid zone of the southern side and most of the eastern side of the Mediterranean. In addition, the expected rise of the sea level by c.40 cm during the twenty-first century (IPCC 2001) will threaten many coastal areas and the millions of people who live there (Chapter 13). A hierarchy of the expected impacts of the drivers linked to global change upon the ten Mediterranean regional hotspots of biodiversity (Médail and Quézel 1997, 1999), was assessed by the group of RICAMARE experts (Table 23.7) (Troumbis et al. 2001). Southern hotspots will be threatened by the extension of arid conditions, whereas northern hotspots will be endangered by an increasing impact of some disturbance events (fires), albeit favoured by the northward extension of Mediterranean-type ecosystems. The sensitivity of habitats will differ according to species assemblages, and a complex pattern of change will occur, depending on location, species composition, and the magnitude of trophic interactions. Saxicolous communities will probably be less affected than grassland, forest, and coastal communities. Fortunately, many rare plant endemics are located in communities with high stress levels (e.g. cliffs, rocks, and screes), which constitute the best refugia against climate changes (Figure 23.19) (see Médail and Diadema 2009, for a recent review). On the other hand, rocky grasslands and steppes could be seriously endangered by the direct ecological consequences of global change including land use changes, aridification processes, increases in the productivity and biomass of

644

Jacques Blondel and Frédéric Médail

TABLE 23.7. Major influences on biodiversity in the Mediterranean hotspots shown in Figure 23.1 according to the group of RICAMARE experts. Where data are available, the relative impact of each factor is given in each case. Note that the Betic-Rifan complex has been divided into two groups in this analysis because of marked contrasts in land use practices. Mediterranean hotspots High and Middle Atlas Rif Mountains Betic region (Andalucia) Maritime and Ligurian Alps Western Mediterranean Islands Southern and Central Greece Crete Southern Anatolia and Cyprus Syria, Lebanon, and Israel Mediterranean Cyrenaica Djurdjura-Kabylies

FIRE

URBA

FORE

GRAZ

AGRI

INVA

N DEP

CO2

UV B

TEMP

RAIN

2 2 3 2 3 3 2 2 3 2 2

0 1 2 2 2 2 1 2 3 2 1

1 2 1 1 1 2 1 1 2 3 2

3 1 1 1 2 2 2 3 1 3 2

1 3 2 1 1 2 2 2 1 2 2

1 1 1 1 2 2 2 1 0 2? 1

? 1 ? ? ? ? ? ? ? ? ?

? ? ? ? ? ? ? ? ? ? ?

2? 1? 1? 2? ? ? ? 2? ? 1? ?

2 2 3 2 2 3 3 1 3 2 3

3 2 3 2 2 3 3 1 3 2 2

Notes: Drivers: FIRE: impact of fires; URBA: urbanization; FORE: forestry practices; GRAZ: grazing; INVA: biological invasions; N DEP: nitrogen deposition; CO2 : CO2 deposition; UV B: ultra-violet deposition; TEMP: putative sensitivity to global warming; RAIN: putative sensitivity to modifications of the rainfall regime. Scale of driver impacts: 0 = no impact; 1 = low impact; 2 = medium impact; 3 = strong impact. Source: Modified from Troumbis et al. 2001. RICAMARE is ‘Research in Global Change in the Mediterranean: a Regional Network.

shrubs and trees, and areal extensions of severe competitive aliens. The question of the extent to which organisms will be able to cope with the new challenges generated by climate change is still largely open, especially

because climate change appears to be taking place at an unprecedented rate in recent human history. In the Mediterranean region there is recent evidence that microevolutionary changes can occur rapidly in fitnessrelated traits such as the breeding season of birds (Visser

Fig. 23.19. Cliffs are important habitats for several rare birds as well as rare endemic plant species throughout the Mediterranean basin (photo: Jacques Blondel).

Biodiversity and Conservation

et al. 2003) or the flowering time in plants (Penuelas et al. 2002). But differential responses of organisms engaged in complex food chains or symbiotic associations may disrupt interactions that are essential for ecosystem functioning such as pollination or seed dispersal. Indeed, different types of responses to global change must be incorporated into management decisions related to conservation and restoration ecology in order to allow endangered populations to adapt themselves to these changing selective pressures (Stockwell et al. 2003). Studying the mechanisms whereby organisms will become adapted to their environment is an issue of major importance for predicting their response to ongoing climate changes. The combination of genetic

and palaeoecological studies has recently provided several interesting insights into the capacity of plants to cope with drastic climatic changes. These studies have also demonstrated the crucial role of glacial refugia for the long-term persistence of the Mediterranean biodiversity (e.g. Tzedakis et al. 2002; Petit et al. 2003). A recent detailed analysis based upon intraspecific phylogeographical studies of eighty-two plant species led to the identification of fifty-two refugia located in the Mediterranean region (Médail and Diadema 2009). These ecologically and climatically stable areas have a high conservation priority because they represent significant reservoirs of endemic plants and of unique genetic diversity for Mediterranean species. But these refugia

Established refugia

(a) Refugia 100 km

645

100 km

Boundary of Mediterranean region

(b) Population density N

Inhabitants per km 2

> 1,000 250 –1,000 100 – 250 50 – 100 0 – 50

Fig. 23.20. Distribution maps within the Mediterranean bioclimatic region (limits indicated by a broken line) of the fifty glacial refugia based upon (a) the phylogeographic structure of eighty-two plant species, and (b) of the human population density, following a 100 × 100 km lattice (after Médail and Diadema 2006).

646

Jacques Blondel and Frédéric Médail

constitute some of the most threatened territories of the Mediterranean basin, owing to higher than average demographic pressures: c.25 per cent of the areas where refugia occur are located in highly populated regions (>250 people per km2 ) (Médail and Diadema 2006, 2009; Figure 23.20). After 10,000 years or more of cohabitation between humans and nature, most Mediterranean ecosystems are so inextricably linked to human interventions that the future of biological diversity cannot be disconnected from that of human affairs, which have too often been characterized by bitter conflicts. Whatever the future of people and their environment, landscape conditions will continue to be shaped and driven by the long common history they share (Foster et al. 2003). Thus, the greatest challenge for the decades to come is to promote sustainable development through the control of population growth and the development of appropriate indicators to assess how environment and development should be integrated. The Mediterranean region, at the crossroads of three continents, between developed countries in the north and developing countries in the south and east, between deserts and fertile lands, has always been and still is an area of both division and convergence. Because it is, in a sense, a microcosm that is representative of many worldwide problems, it is an exceptional laboratory and pilot region for launching a strategy for sustainable development on a regional scale.

Acknowledgements We are deeply indebted to Jamie Woodward who invited us to write this chapter. We also thank Oxford University Press for giving us this opportunity to make a point on the state of biodiversity in this fascinating region. Jamie Woodward and Kathy Willis much improved a first draft of the manuscript. Thanks to them.

References Affre, L., Thompson, J. D., and Debussche, M. (1997), Genetic structure of continental and island populations of the Mediterranean endemic Cyclamen balearicum (Primulaceae). American Journal of Botany 84: 437–51. Andrieu-Ponel, V., Ponel, P., Jull, A. J. T., de Beaulieu, J. L., Bruneton, H., and Leveau, P. (2000), Towards the reconstruction of the Holocene vegetation history of Lower Provence: two new pollen profiles from Marais des Baux. Vegetation History and Archaeobotany 9: 71–84. Araùjo, M. B. (2003), The coincidence of people and biodiversity in Europe. Global Ecology and Biogeography 12: 5–12. Attenborough, D. (1987), The First Eden. The Mediterranean World and Man. Fontana/Collins, London.

Balmford, A., Green, R. E., and Jenkins, M. (2003), Measuring the changing state of nature. Trends in Ecology and Evolution 18: 326–30. Bañares, A., Blanca, G., Güemes, J., Moreno, J. C., and Ortiz, S. (eds.) (2003), Atlas y libro rojo de la flora vascular amenazada de España. Dirección General de Conservación de la Naturaleza, Madrid. Barbero, M., Bonin, G., Loisel, R., and Quézel, P. (1990), Changes and disturbances of forest ecosystems caused by human activities in the western part of the Mediterranean basin. Vegetatio 87: 151–73. Beaulieu, J. L. de, Miras, Y., Andrieu-Ponel, V., and Guiter, F. (2005), Vegetation dynamics in North-Western Mediterranean regions: instability of the Mediterranean bioclimate. Plant Biosystems 139: 114–26. Benoit, G. and Comeau, A. (2005), Méditerranée. Les perspectives du Plan Bleu sur l’environnement et le développement. Éditions de l’Aube et Plan Bleu, La Tour d’Aigues. BirdLife International (2000), Threatened Birds of the World. Lynx Edicions, Barcelona, and BirdLife International, Cambridge. (2004) Birds in Europe: Population Estimates, Trends and Conservation Status. BirdLife International, Cambridge. Blondel, J. (1987), From biogeography to life history theory: a multithematic approach. Journal of Biogeography 14: 405–22. (2006), The ‘design’ of Mediterranean landscapes: A millennial story of humans and ecological systems during the historic period. Journal Human Ecology 34: 713–29. (2008), On humans and wildlife in Mediterranean islands. Journal of Biogeography 35: 509–18. and Aronson, J. (1995), Biodiversity and ecosystem function in the Mediterranean basin: Human and non-human determinants, in G. Davis and D. M. Richardson (eds.), Biodiversity and Ecosystem Function in Mediterranean-type Ecosystems. Ecological Studies 109, Springer, Berlin, 43–119. (1999), Biology and Wildlife of the Mediterranean Region. Oxford University Press, Oxford. and Farré, H. (1988), The convergent trajectories of bird communities in European forests. Oecologia 75: 83–93. and Vigne, J.-D. (1993), Space, time and man as determinants of vertebrate diversities in the Mediterranean region, in R. E. Ricklefs and D. Schluter (eds.), Historical and Geographical Determinants of Community Diversity. Chicago University Press, Chicago, 135–46. Dias, P. C., Perret, P., Maistre, M., and Lambrechts, M. M. (1999), Selection-based biodiversity at a small spatial scale in an insular bird. Science 285: 1399–402. ——Dias, P. C., and Lambrechts, M. M. (2001), Is phenotypic variation of Blue tits (Parus caeruleus L.) in Mediterranean mainland and insular landscapes adaptive? Genetics, Selection, Evolution 33 (suppl. 1): S121–S139. Boydak, M. (1987), A new occurrence of Phoenix theophrasti in Kumluca-Karaöz, Turkey. Principes 31: 89–95. Braudel, F. (1986), La Méditerranée, les Hommes et l’Héritage. Flammarion, Paris. Brosset, A. (1990), L’Évolution récente de l’avifaune du nord-est marocain: pertes et gains depuis 35 ans. Revue d’Écologie (Terre et Vie) 45: 237–44. Cadi, A. and Joly, P. (2004), Impact of the introduction of the red-eared slider (Trachemys scripta elegans) on survival of the European pond turtle (Emys orbicularis). Biodiversity and Conservation 13: 2511–18. Catsadorakis, G. (2003), Greece’s Heritage from Nature. WWF, Athens.

Biodiversity and Conservation Charrier, A. (1995), France maintains strong tradition of support for biodiversity activities worldwide. Diversity 11: 89–90. Chauvet, J.-M., Brunel Deschamps, E., and Hillaire, C. (1995), La Grotte Chauvet. Seuil, Paris. Cincotta, R. P. and Engelman, R. (2000), Nature’s Place. Human Population and the Future of Biological Diversity. Population Action International, Washington DC. Colas, B., Olivieri, I., and Riba, M. (1997), Centaurea corymbosa, a cliff-dwelling species tottering on the brink of extinction: a demographic and genetic study. Proceedings of the National Academy of Sciences (USA) 94: 3471–6. (2001), Spatio-temporal variation of reproductive success and conservation of the narrow-endemic Centaurea corymbosa (Asteraceae). Biological Conservation 99: 375–86. Collins, N. M. and Thomas, J. A. (1991), The Conservation of Insects and their Habitats. Academic Press, London. Crivelli, A. J. (1996), The Freshwater Fish Endemic to the Northern Mediterranean Region. Tour du Valat, Arles. and Maitland, P. S. (1995), Endemic freshwater fishes of the northern Mediterranean region. Biological Conservation 72: 1–337. Debussche, M., Debussche, G., and Lepart, J. (2001), Changes in the vegetation of Quercus pubescens woodland after cessation of coppicing and grazing. Journal of Vegetation Science 12: 81–92. Décamps, H. and Décamps, O. (2002), Mediterranean Riparian Woodlands. Conservation of Mediterranean Wetlands. Medwet series 12. Tour du Valat, Arles. Delanoë, O., de Montmollin, B., and Olivier, L. (1996), Conservation of Mediterranean Island Plants. 1. Strategy for Action. IUCN, Gland and Cambridge. De Stefano, L. (2004), Freshwater and Tourism in the Mediterranean. WWF Mediterranean Programme, Rome. Diadema, K., Médail, F., and Bretagnolle, F. (2007), Fire as a control agent of demographic structure and plant performance of a rare Mediterranean endemic geophyte. Comptes-Rendus Biologies 30: 691–700. Diamond, J. M. (2002), Evolution, consequences, and future of plant and animal domestication. Nature 418: 700–7. Dias, P. C., Verheyen, G. R., and Raymond, M. (1996), Sourcesink populations in Mediterranean blue tits: evidence using single-locus minisatellite probes. Journal of Evolutionary Biology 9: 965–78. di Castri, F., Hansen, A. J., and Debussche, M. (eds.) (1990), Biological Invasions in Europe and the Mediterranean Basin. Kluwer, Dordrecht. Donald, P. F., Green, R. E., and Heath, M. F. (2001), Agricultural intensification and the collapse of Europe’s farmland bird populations. Proceedings of the Royal Society of London B 268: 25–9. Drake, J. A., Mooney, H. A., di Castri, F., Groves, R. H., Kruger, F. J., Rejmanek, M., and Williamson, M. (1989), Biological Invasions. A Global Perspective. John Wiley & Sons, Chichester. Dupouey, J. L., Dambrine, E., Moares, C., and Lafitte, J. D. (2002), Irreversible impact of past land use on forest biodiversity. Ecology 83: 2978–84. Entwistle A., Atay, S., Byfield, A., and Oldfield, S. (2002), Alternatives for the bulb trade from Turkey: a case study of indigenous bulb propagation. Oryx 36: 333–41. Eugenio, M. and Lloret, F. (2006), Effects of repeated burning on Mediterranean communities of the northeastern Iberian Peninsula. Journal of Vegetation Science 17: 755–64.

647

Fenech, N. (1993), Fatal Flight: The Maltese Obsession with Killing Birds. Quiller, Fakenham. Foster, D. R. (2002), Conservation issues and approaches for dynamic cultural landscapes. Journal of Biogeography 29: 1533–5. Schoonmaker, P. K., and Pickett, S. T. A. (1990), Insights from paleoecology to community ecology. Trends in Ecology and Evolution 5: 119–22. Motzkin, G. M., Bernardos, D., and Cardoza, J. (2002), Wildlife dynamics in the changing New England landscape. Journal of Biogeography 29: 1319–57. Swanson, F., Aber, J., Burke, I., Brokaw, N., Tilman, D., and Knapp, A. (2003), The importance of land use and its legacies to ecology and environmental management. BioScience 53: 77–88. Georgoudis, A. (1995), Animal genetic diversity plays important role in Mediterranean agriculture. Diversity 11: 16–19. Goren-Inbar, N., Alperson, N., Kislev, M. E., Simchoni, O., Melamed, Y., Ben-Nun, A., and Werker, E. (2004), Evidence of hominin control of fire at Gesher Benot Ya’aqov, Israel. Science 304: 72–7. Gramond, D. (2002), Dynamique de l’occupation du sol et variation des usages de l’eau en Anatolie centrale (Turquie) au cours du XXe siècle. Recherches méthodologiques basées sur l’analyse diachronique de données satellites et statistiques. D.Phil. thesis, University of Paris IV, Sorbonne. Granval, P. and Muys, B. (1992), Management of forest soils and earthworms to improve woodcock (Scolopax sp.) habitats: a literature survey. Gibier et Faune Sauvage 9: 243–55. Grenon, M. and Batisse, M. (1989), Futures for the Mediterranean Basin: The Blue Plan Oxford University Press, Oxford. Greuter, W. (1994), Extinctions in Mediterranean areas. Philosophical Transactions of the Royal Society of London B 344: 41–6. Grillas, P., Gauthier, P., Yavercovski, N., and Perennou, C. (2004), Mediterranean Temporary Pools, i. Issues Relating to Conservation, Functioning and Management. Tour du Valat, Arles. Grove, A. T. and Rackham, O. (2001), The Nature of Mediterranean Europe. An Ecological History. Yale University Press, New Haven. Gurevitch, J. and Padilla, D. K. (2004), Are invasive species a major cause of extinction? Trends in Ecology and Evolution 19: 470–4. Huston, M. A. (1994, Biological Diversity: The Coexistence of Species on Changing Landscapes. Cambridge University Press, Cambridge. IPCC (2001), Third Assessment Report: Climate Change 2001. Cambridge University Press, Cambridge. Jump, A. S. and Peñuelas, J. (2005), Running to stand still: adaptation and the response of plants to rapid climate change. Ecology Letters 8: 1010–20. Kiss, L. and Magnin, F. (2006), High resilience of Mediterranean land snail communities to wildfires. Biodiversity and Conservation 15: 2925–44. and Torre, F. (2004), The role of landscape history and persistent biogeographical patterns in shaping the responses of Mediterranean land snail communities to recent fire disturbances. Journal of Biogeography 31: 145–57. Kuhnholtz-Lordat, G. (1938), La Terre incendiée. Essai d’agronomie comparée. La Maison Carrée, Nîmes. Laguna, E., Deltoro, V. I., Pèrez-Botella, J., Pèrez-Rovira, P., Serra, L., Olivares, A., and Fabregat, C. (2004), The role of small

648

Jacques Blondel and Frédéric Médail

reserves in plant conservation in a region of high diversity in eastern Spain. Biological Conservation 119: 421–6. Le Houérou, H. N. (1981), Impact of man and his animals on Mediterranean vegetation, in F. Di Castri, D.W. Goodall, and R.L. Specht (eds.), Ecosystems of the World, xi. MediterraneanType Shrublands. Elsevier, Amsterdam, 479–521. (1990), Global change: vegetation, ecosystems, and land use in the southern Mediterranean basin by the mid twentyfirst century. Israel Journal of Botany 39: 481–508. Leon, C., Lucas, G., and Synge, H. (1985), The value of information in saving threatened Mediterranean plants, in C. Gomez-Campo (ed.), Plant Conservation in the Mediterranean Area. Dr W. Junk, Dordrecht, 177–96. Liven-Schulman, I., Leshem, Y., Alon, D., and Yom-Tov, Y. (2004), Causes of population declines of the Lesser Kestrel Falco naumanni in Israel. Ibis 146: 145–52. Lloret, F., Médail, F., Brundu, G., Mus, M., and Hulme, P. E. (2004), Local and regional abundance of exotic plant species on Mediterranean islands: are species traits important? Global Ecology and Biogeography 13: 37–45. McNeill, J. R. (1992), The Mountains of the Mediterranean World, an Environmental History. Cambridge University Press, Cambridge. Magnin, G. (1991), Hunting and persecution of migratory birds in the Mediterranean region, in T. Salathé (ed.), Conserving Migratory Birds. International Council for Bird Preservation, Cambridge, 59–71. Marinkovic, S. and Karadzic, B. (1996), Role of nomadic farming in the distribution of Griffon Vulture Gyps fulvus on the Balkan peninsula, in Hellenic Zoological Society (ed.), Seventh International Conference on the Zoogeography of Greece and Adjacent Countries. Athens, 141–52. Martin, P. S. (1984), Prehistoric overkill: the global model, in Martin and Klein (1984: 354–403). and Klein, R. G. (eds.) (1984), Quaternary Extinctions. University of Arizona Press, Tucson. Mazzoleni, S., Di Pascale, G., Di Martino, O., Rego, F., and Mulligan, M. (eds.) (2004), Recent Dynamics of Mediterranean Vegetation and Landscape. John Wiley & Sons, London. Médail, F. and Diadema, K. (2006), Biodiversité végétale méditerranéenne et anthropisation: approches macro et microrégionales. Annales de Géographie, special issue: Les territoires de la biodiversité 651: 618–49. (2009), Glacial refugia influence plant diversity patterns in the Mediterranean Basin. Journal of Biogeography (published online on January 26, 2009). and Myers, N. (2004), Mediterranean basin, in Mittermeier et al. (2004: 144–7). and Quézel, P. (1997), Hot-spots analysis for conservation of plant biodiversity in the Mediterranean basin. Annals of the Missouri Botanical Garden 84: 112–27. (1999), Biodiversity hotspots in the Mediterranean basin: setting global conservation priorities. Conservation Biology 13: 1510–13. Roche, P., and Tatoni, T. (1998), Functional groups in phytoecology: an application to the study of isolated plant communities in Mediterranean France. Acta Oecologica 19: 263–74. Mittermeier, R. A., Robles Gil, P., Hoffmann, M., Pilgrim, J., Brooks, T., Mittermeier, C. G., Lamoreaux, J., and da Fonseca, G. A. B. (eds.) (2004), Hotspots Revisited: Earth’s Biologically Richest and Most Endangered Terrestrial Ecoregions.

CEMEX, Monterrey, Conservation International, Washington, and Agrupación Sierra Madre, Mexico. Mooney, H. A. and Hobbs, R. J. (2000), Invasive Species in a Changing World. Island, Washington DC. Moreno, J. M. and Oechel, W. C. (1994), The Role of Fire in Mediterranean-type Ecosystems. Springer, New York. Muntaner, J. and Mayol, J. (1996), Biology and Conservation of Mediterranean Raptors. SEO/Birdlife, Palma de Mallorca. Natali, A. and Jeanmonod, D. (1996), Flore analytique des plantes introduites en Corse, in D. Jeanmonod and H. M. Burdet (eds.), Compléments au prodrome de la flore corse. Conservatoire et Jardin botaniques de Genève, Geneva. Naveh, Z. (1999), The role of vegetation as an evolutionary and ecological factor on the landscapes and vegetation of Mt. Carmel. Journal of Mediterranean Ecology 1: 11–25. and Dan, J. (1973), The human degradation of Mediterranean landscapes in Israël, in F. di Castri and H. A. Mooney (eds.), Mediterranean Type Ecosystems. Origin and Structure. Springer, Heidelberg, 372–90. Noss, R. F. and Csuti, B. (1994), Habitat fragmentation, in G. K. Meffe and R. C. Carroll (eds.), Principles of Conservation Biology. Sinauer, Sunderland, Mass., 237–64. Oldfield, S., Lusty, C., and McKinven, A. (1998), The World List of Threatened Trees. IUCN, World Conservation Press, Cambridge. Pearce, F. and Crivelli, A. J. (1994), Characteristics of Mediterranean Wetlands. Medwet series 1. Tour du Valat, Arles. Penuelas, J., Filella, I., and Comas, P. (2002), Changed plant and animal life cycles from 1952 to 2000 in the Mediterranean region. Global Change Biology 8: 531–44. Petit, R. J., Aguinagalde, I., de Beaulieu, J. L., Bittkau, C., Brewer, S., Cheddadi, R., Ennos, R., Fineschi, S., Grivet, D., Lascoux, M., Mohanty, A., Müller-Starck, G., Demesure-Musch, B., Palmé, A., Martin, J. P., Rendell, S., and Vendramin, G. G. (2003), Glacial refugia: hotspots but not melting pots of genetic diversity. Science 300: 1563–5. Piñol, J., Terradas, J., and Lloret, F. (1998), Climate warming, wildfire hazard, and wildfire occurrence in coastal Eastern Spain. Climate Change 38: 345–57. Plantlife International and IUCN (2003), Defining Important Plant Areas in the Mediterranean Region. Workshop Report. IUCN Centre for Mediterranean Cooperation, Malaga. Pons, A. and Quézel, P. (1985), The history of the flora and vegetation and past and present human disturbance in the Mediterranean region, in C. Gomez-Campo (ed.), Plant Conservation in the Mediterranean Area. Dr W. Junk, Dordrecht, 25–43. and Thinon, M. (1987), The role of fire from paleoecological data. Ecologia Mediterranea 13: 3–11. Prodon, R., Fons, R., and Athias-Binche, F. (1987), The impact of fire on animal communities in Mediterranean area, in L. Trabaud (ed.), The Role of Fire in Ecological Systems. SPB Academic, The Hague, 121–57. Pullin, A. S. (1995), Ecology and Conservation of Butterflies. Chapman & Hall, London. Quézel, P. (2000), Réflexions sur l’évolution de la flore et de la végétation au Maghreb méditerranéen. Ibis, Paris. and Médail, F. (2003), Écologie et biogéographie des forêts du bassin méditerranéen. Elsevier, Paris. Barbero, M., Bonin, G., and Loisel, R. (1990), Recent plant invasions in the circum-mediterranean region, in di Castri, Hansen, and Debussche (eds.) (2004: 51–60).

Biodiversity and Conservation Raimondo, F. M., Gianguzzi, L., and Ilardi, V. (1994), Inventario delle specie ‘a rischio’ nella flora vascolare nativa della Sicilia. Quadernari Botanica Ambientale Applicata 3: 65–132. Ramade, F. (1997), Conservation of Mediterranean Ecosystems. 2nd edn. Les Fascicules du Plan Bleu 3. Economica, Paris. Rhazi, L., Grillas, P., Toure, A. M., and Tan Ham, L. (2001), Impact of land use in catchment and human activities on water, sediment and vegetation of Mediterranean temporary pools. Compte-Rendus de l’Académie des Sciences de Paris, Sciences de la vie/Life Sciences 324: 165–77. Charpentier, A., and Médail, F. (2004), Experimental management of Mediterranean temporary pools for conservation of the rare quillwort Isoetes setacea. Biological Conservation 118: 675–84. Roi, J. (1937), Les espèces eurasiatiques continentales et les espèces boréo-alpines dans la région méditerranéenne occidentale. Ph.D. thesis, Université de Montpellier, Imprimerie Fournié, Toulouse. Sala, O. E., Chapin III, F. S., Armesto, J. J., Berlow, E., Bloomfield, J., Dirzo, R., Huber- Sanwald, E., Huenneke, L. F., Jackson, R. B., Kinzig, A., Leemans, R., Lodge, D. M., Mooney, H. A., Oesterheld, M., Leroy Poff, N., Sykes, M. T., Walker, B. H., Walker, M., and Wall, D. H. (2000), Global biodiversity scenarios for the year 2100. Science 287: 1770–4. Santos, T. and Tellería, J. L. (1994), Influence of forest fragmentation on seed consumption and dispersal of Spanish Juniper Juniperus thurifera. Biological Conservation 70: 129–34. (1997), Vertebrate predation on Holm Oak, Quercus ilex, acorns in a fragmented habitat: effects on seedling recruitment. Forest Ecology and Management 98: 181–7. Sezik, E. (1989), Turkish orchids and salep, in Société Française d’Orchidophilie (ed.), Orchidées botaniques du monde entier. Société Française d’Orchidophilie, Paris, 181–9. Simmons, A. H. (1988), Extinct pygmy hippopotamus and early man in Cyprus. Nature 333: 554–7. (1991), Humans, island colonization, and Pleistocene extinctions in the Mediterranean: the view from Akrotiri Aetokremnos, Cyprus. Antiquity 65: 857–69. Sket, B. (1999), The nature of biodiversity in hypogean waters and how it is endangered. Biodiversity and Conservation 8: 1319–38. Skinner, J. and Zalewski, S. (1995), Fonctions et valeurs des zones humides méditerranéennes. Medwet Séries 2. Tour du Valat, Arles. Socias y Company, R. (1990), Breeding self-compatible almonds. Plant Breeding Review 8: 313–38. Speight, M. C. D. (1989), Saproxylic invertebrates and their conservation. Nature and Environment Series 42. Council of Europe, Strasbourg. Stockwell, C. A., Hendry, A. P., and Kinnison, M. T. (2003), Contemporary evolution meets conservation biology. Trends in Ecology and Evolution 18: 94–101. Suehs, C. M., Affre, L., and Médail, F. (2004a), Invasion dynamics of two alien Carpobrotus (Aizoaceae) taxa on a Mediterranean island: I. Genetic diversity and introgression. Heredity 92: 31–40. (2004b), Invasion dynamics of two alien Carpobrotus (Aizoaceae) taxa on a Mediterranean island: II. Reproductive strategies. Heredity 92: 550–6. Médail, F., and Affre, L. (2001), Ecological and genetic features of the invasion by the alien Carpobrotus plants in

649

Mediterranean island habitats, in G. Brundu, J. Brock, I. Camarda, L. Child, and M. Wade (eds.), Plant Invasions: Species Ecology and Ecosystem Management. Backhuys, Leiden, 145–58. Tamisier, A. and Grillas, P. (1994), A review of habitat changes in the Camargue: an assessment of the effects of the loss of biological diversity on the wintering waterfowl community. Biological Conservation 70: 39–47. Teplitsky, C., Plénet, S., and Joly, P. (2003), Tadpoles’ responses to risk of fish introduction. Oecologia 134: 270–7. Terrasse, M. (1996), Réintroduction de rapaces dans l’aire Méditerranéenne de répartition, in J. Muntaner and J. Mayol (eds.), Biology and Conservation of Mediterranean Raptors. SEO/Birdlife, Palma de Mallorca, 251–9. Thirgood, J. V. (1981), Man and the Mediterranean Forest. A History of Resource Depletion. Academic Press, London. Thompson, J. D. (2005), Plant Evolution in the Mediterranean. Oxford University Press, Oxford. Torre, I. and Diaz, M. (2004), Small mammal abundance in Mediterranean post-fire habitats: a role for predators? Acta Oecologica 25: 137–43. Trabaud, L. and Prodon, R. (eds.) (2002), Fire and Biological Processes. Backhuys, Leiden. Troumbis, A., Moreno, J. M., and Médail, F. (eds.) (2001), Report on a RICAMARE Workshop, Biodiversity and Conservation in Changing Mediterranean Landscapes. Lesvos, University of the Aegean, Mytilini. Tucker, G. M., and Heath, M. F. (1994), Birds in Europe: Their Conservation Status. BirdLife Conservation series 3. BirdLife International, Cambridge. Tzedakis, P. C., Lawson, I. T., Frogley, M. R., Hewitt, G. M., and Preece, R. C. (2002), Buffered tree population changes in a Quaternary refugium: evolutionary implications. Science 297: 2044–7. Vallauri, D. (ed.) (2003), Livre blanc sur la protection des forêts naturelles en France; forêts métropolitaines. Tec & Doc, Paris. Vavilov, N. I. (ed.) (1935), Origin and Geography of Cultivated Plants. Nauka, Leningrad. Véla, E. and Benhouhou, S. (2007), Évaluation d’un nouveau point chaud de biodiversité végétale dans le bassin méditerranéen (Afrique du Nord). Comptes-Rendus Biologies 30: 589– 605. Verlaque, R., Médail, F., and Aboucaya, A. (2001), Valeur prédictive des types biologiques pour la conservation de la flore méditerranéenne. Comptes-Rendus de l’Académie des Sciences, Sciences de la Vie/Life Sciences 324: 1157– 65. Vigne, J. D., Bailon, S., and Cuisin, J. (1997), Biostratigraphy of amphibians, reptiles, birds and mammals in Corsica and the role of man in the Holocene faunal turnover. Anthropozoologica 25–6: 587–604. Visser, M. E., Adriaensen, F., Van Balen, J. H., Blondel, J., Dhondt, A. A., Van Dongen, S., Du Feu, C., Ivankina, E. V., Kerimov, A., De Laet, J., Matthysen, E., McCleery, R., Orell, M., and Thomson, D. (2003), Variable responses to large-scale climate change in European Parus populations. Proceedings of the Royal Society London B 270: 367–72. Vitousek, P. M. (1994), Beyond global warming: ecology and global change. Ecology 75: 1861–76. Walter, K. S. and Gillett, H. J. (eds.) (1998), 1997 IUCN Red List of Threatened Plants. Compiled by the World Conservation

650

Jacques Blondel and Frédéric Médail

Monitoring Centre. IUCN-The World Conservation Union, Gland and Cambridge. Williamson, M. (1996), Biological Invasions. Chapman & Hall, London. Woodward, J. (1995), Patterns of erosion and suspended sediment yield in Mediterranean river basins, in I. D. L. Foster, A. M. Gurnell, and B. W. Webb (eds.), Sediment and Water

Quality in River Catchments. John Wiley & Sons, London, 365–89. WWF (World Wildlife Fund) (2001), The Mediterranean Forests, a New Conservation Strategy. WWF Mediterranean Programme, Rome. Zohary, D. and Hopf, M. (2000), Domestication of Plants in the Old World. Oxford University Press, Oxford.

This chapter should be cited as follows Blondel, J. and Médail, F. (2009), Biodiversity and conservation, in J. C. Woodward (ed.), The Physical Geography of the Mediterranean. Oxford University Press, Oxford, 615–650.

Index Bold numbers denote reference to illustrations or tables and their respective captions abstraction 79 groundwater 571, 590, 591 see also aquifers lakes 260, 274, 276, 278, 279, 281 rivers 241, 345, 347, 523 Adriatic Sea 11, 16, 33, 35, 37, 39, 42, 44, 51, 52, 53, 56, 57, 75, 77, 78, 84, 103, 197, 290, 301, 368–9, 385, 387, 388, 392, 401, 406, 472, 496, 497, 508, 515, 516, 522, 531–2, 600, 622, 635, 636 Aegean arc 10, 13, 456, 473, 474 Aegean Sea 8–13, 15, 18, 21–2, 33, 34, 35, 37–9, 42, 44, 51, 52–3, 56, 57, 74, 75, 78, 103, 123, 154, 245, 246, 269, 338, 347, 385, 387, 388, 390, 398, 399, 400, 401, 403, 406, 421, 435, 436, 450–1, 454, 456, 459, 470–1, 472, 478–80, 482–3, 485, 493, 496, 499, 500, 503, 504, 508, 509–510, 515, 534, 566, 579, 603 Aegean volcanic province 436, 450–57 see also volcanoes and volcanism aerosols 415–7, 418–19, 437, 439–40, 449–50, 529, 599–601, 607–11 see also aeolian processes and landforms; TOMS aerosol haze see air quality Aeolian Islands 436, 446–9 see also volcanoes and volcanism aeolian processes and landforms 415–25 dunes and other landforms 325, 386, 405, 422–5, 571 dust sources and trajectories 58, 75, 119, 175, 415–22, 599 see also TOMS; loess 420–2 loess 176, 304, 415, 416, 420–2, 425 see also aeolian processes and landforms afforestation 208, 249, 571 see also reafforestation African plate 8, 11, 36, 324, 435, 450, 469 see also tectonics and landscape development aftershocks see and seismicity earthquakes aggradation (fluvial) 17–18, 322, 323, 327, 329, 331–5, 338, 339 see also river systems and environmental change

agriculture see also pastoralism; irrigation; land degradation fertility and nutrients 573, impacts 182, 184, 192, 239, 272, 571, 586, 603, 609, 617, 621 see also soils land use change 224–5, 250, 263, 272, 276, 301, 314, 406, 423, 572, 575, 621, 625, 633 mechanization and intensification 179, 564, 618, 637–8 origins and early expansion 213–4, 222–3, 271, 338, 546, 617, 638 ploughing 182, 222, 239, 249, 336 see also tillage pollution 274, 280, 591 see also eutrophication storm damage 534, 535 volcanic hazards and agriculture 437, 438–9, 455 agricultural terraces 214, 239, 301, 303, 337, 591 abandonment 214, 224, 549 see also land abandonment Aguas River, Spain 24, 25, 26, 245, 328, 330, 333, 334 air masses 37, 39, 42, 44, 47, 48, 55, 73, 513–4, 519, 537, 601–3, 605 see also climate air quality 437, 440, 553, 561, 599–600, 603, 605, 609, 610–611 aerosol haze 437, 449, 601, 608–10 see also aeolian processes and landforms climatic controls on 600–3 climate and the water cycle 610–11 oxidation capacity 607–8 ozone dynamics 604–6 upper tropospheric pollution plume 603–4 Albania 11, 75, 233, 236, 256, 257, 263, 266, 275, 277, 355, 357, 365, 367, 378, 386, 448, 495, 497, 508, 541, 587, 588, 594, 596, 631, 640 Alboran basin 9, 10, 385, 387, 389 Alboran Sea 10, 13, 39, 40, 41, 42–3, 58, 103, 105, 120, 121, 122, 377, 386, 388, 495–6, 508, 510, 532 Alcantara River, Spain 335

Aleppo pine 206, 207, 210, 220, 237, 551 Alexander the Great 485 Alexandria 40, 69, 70, 71, 386, 389, 418, 424, 471, 498, 499 Alfios River, Greece 239, 320, 321, 346 Algarve 211 Algeria 10, 19, 20, 39, 40, 41, 42, 58, 78, 108, 116, 150, 197, 232, 233, 236, 244, 247, 274, 275, 287, 315, 319, 376, 386, 416, 417, 418, 424, 469, 470, 480, 481, 482, 485, 495, 496, 513, 514, 515, 516, 517, 532, 533, 535, 541, 571, 587, 589, 590, 591, 592, 594, 595, 596, 618, 626, 628, 631, 640 Algerian basin 7, 9, 10, 41, 385, 387, 388 Algerian current 40, 41 Algiers 41, 71, 74, 495, 536 Alkyonides Gulf 22–3 allopatric speciation 140, 148, 160, 216 alluvial archaeology 319 see also geoarchaeology alluvial fans 7, 18, 23, 177, 178, 325, 359 alluvial history see river systems and environmental change alluvial soils 234, 332, 177, 178 Almanzora River, Spain 249 Almeria, Spain 7, 19, 40, 41, 70, 109, 187, 245, 567 Alpine-Himalayan belt 5, 6 Alps (Austrian/French/Swiss) 10, 11, 12, 38, 70, 74, 77, 84, 113, 115, 124, 141, 143, 149, 169, 174, 185, 190, 191, 195, 255, 263, 293, 301, 325, 331, 344, 361, 370, 373, 376, 377, 417, 418, 421, 515, 521, 531, 585, 600, 603 Alps-Betics orogen 7, 10 AMS radiocarbon dating see radiocarbon dating Anatolia 8, 11, 75, 86, 116, 139, 141, 144, 156, 169, 171, 177, 207, 208, 215, 222, 255, 261, 262, 263, 385, 390, 403, 469, 515, 583, 616, 621, 644 Anatolian Fault, North (NAF) 470, 472–3, 474, 477, 478, 480

652

Index

Anatolian plate 473 Ancient Greece 469 see also Classical Period Andalucia 214, 264, 265, 516, 521, 567, 572, 630, 644 Andorra 360, 516, 522 anoxia 269 see also sapropels Antarctica 73 Antarctic gateway 91 Antarctic ice cores 124 Antarctic ice sheets 91, 92, 95, 126 anticyclones 47, 74, 77, 78, 126, 127, 418, 519, 521, 531, 537, 602, 603, 604, 605, 610 see also cyclones and cyclogenesis; air masses; climate aphelion 45, 46 Appenines, Italy 16, 357, 369–370 Apulia, Italy 10, 139, 288, 295, 385, 400, 401, 508 aquifers 196, 251, 274, 276, 281, 287, 295–8, 299, 307, 309, 313, 571, 583, 584, 586, 590, 591, 592 see also water resources Arabian Desert 415, 418 Arabian Peninsula 602, 603 Arabian plate 36, 144, 147, 324 see also tectonics and landscape development Arabian Sea 104 archaeology 140, 281, 319, 337 see also geoarchaeology architecture 306, 313, 315, 485 Argive plain, Greece 338 aridification 93, 94, 95, 643 see also aridity; aeolian processes and landforms aridity 47, 57, 93, 97, 99, 100, 101, 102, 107, 114, 116, 119, 123, 126, 127, 169, 171, 178, 204, 209, 211, 234, 295, 376, 422, 568, 583, 585, 621, 643 Arles, France 344, Arno River, Italy 345, 388, 404, 621 aromatic plants 204, 206 artefacts 476 see also coins; pottery; archaeology; stone tools Asia 69, 73, 75, 93, 107, 112, 113, 139, 141, 142, 144, 145, 146, 147, 155, 209, 216, 229, 425, 483, 510, 514, 593, 602, 603, 604, 606, 610, 629 see also Eurasia Asia Minor 176, 469, 522, 593 Asia Minor Current 41 Asian monsoon 39, 46, 55, 75, 604 see also Indian monsoon astronomical parameters see Milankovitch cycles Aswan High Dam 39, 40, 97, 232, 272, 405, 585, 590 see also Nile River Athens 69, 70, 71, 470, 476, 479, 484, 503, 514, 523, 535

Atlantic Ocean see North Atlantic Atlas Mountains 21, 58, 70, 75, 108, 116, 123, 149, 150, 154, 177, 207, 212, 215, 255, 261, 269, 272, 273, 287, 303, 354, 360, 361, 375–376, 416, 517, 616, 628, 644 atmospheric instability see cyclones and cyclogenesis; storms and floods; climate Australia 69, 157, 305, 541, 546, 547, 579, 615, 629 Austria 11, 74, 301 Axios River, Greece 245, 246, 347, 388, 621 Azores 78, 440 Azores High 37, 386, 601–602 back-arc basins 7, 8, 9, 10, 11, 13, 331 see also tectonics and landscape development badlands 5, 7, 18–20, 24, 25, 26, 185, 187, 239, 320, 334, 570 Bagni de Tivoli, Italy 306 Balearic basin 20, 43, 385, 387, 388, 400 Balearic Islands 10, 423, 424, 496, 513, 514, 515, 516, 532, 533, 536, 623, 631 see also Majorca Balkan Peninsula 102, 111, 139, 141, 142, 233, 287, 301, 513, 515, 525, 584, 600, 638 Balkans 74, 75, 78, 80, 111, 143, 144, 149, 156, 159, 176, 223, 233, 255, 261, 264, 266–7, 274, 331, 355–7, 362, 365–9, 470, 522, 596, 635 Barcelona 86, 282, 360, 516, 519, 522, 532, 533, 534, 572 basalt 14, 16, 171, 335 baseflow 586 see also runoff base level 7, 15, 16, 18, 19, 23, 24, 25, 26, 287, 293, 296, 301, 304, 307, 325, 334–5, 389, 480, 483 see also tectonics and landscape development Basento River, Italy 388 basin and range environments 7, 17 see also tectonics and landscape development beaches see coastal environments bed load see sediment loads and yields bedrock erosion and weathering 17, 171, 173, 175, 185, 189, 235, 290, 292, 298, 302, 303, 306, 311, 325, 327, 357, 378, 395, 484 Benioff zone 8, 13 Betic Cordillera, Spain 5, 7, 10, 11, 24, 293, 333, 616 Betic corridor 7, 14, 36, 94, 389 see also Rifian corridor bioclimate and bioclimatic zones 127, 149, 151, 203, 205, 630, 640, 645

bioconstruction 394, 396, 397 biodiversity 140, 150, 157–60, 203, 206, 214–7, 224, 225, 250, 258, 266, 272, 274, 277, 279, 281, 282, 396, 406, 596, 615–646 see also species diversity biodiversity and conservation 615–646 see also conservation biodiversity decline 629–40 conservation frameworks 640–3 fire and biodiversity 624–6 global change and biodiversity 643–6 habitat destruction and fragmentation 618–24 hotspots 615–6, 644 human impact on 616–29 invasive species 626–9 biogeography 5, 142, 149, 155, 210 biomass 93, 123, 158, 173, 209, 216, 396, 546, 556, 573, 574, 575, 576, 579, 603, 609, 643 biostratigraphy 401, 498 see also bioconstruction; pollen and pollen records; foraminifera; diatoms bioturbation 36, 37, 182, 267 bird fauna 125, 140, 141, 142, 143, 147–8, 151, 152, 154, 155, 156, 214, 266, 274, 276, 281, 618, 621, 623, 624, 625, 628, 633, 634, 636, 637, 638, 641, 642, 644 BirdLife International 641, 642 Black Sea 10, 15, 23, 39, 44, 74, 75, 108, 247, 255, 274, 275, 355, 402, 522, 604, 609 black shales 33, 36 blue-green algae 267 Blue Plan 232 Bora wind 35, 37, 44, 74, 77, 78, 513, 531, 537 see also climate; storms and floods boreal summer 45, 46, 54, 99, 100, 101, 102, 116, 117, 123, 127 boreal winter 45, 46, 54 Bosnia-Herzegovina 11, 275, 288, 353, 355, 541, 587, 588, 594, 640 boulder bed channels 309, 324, 343, 344 see also gravel bed rivers braided rivers 18, 345, 404, 422 see also gravel bed rivers breccia 456 Bronze Age 18, 147, 158, 185, 192, 251, 269, 270, 271, 315, 336, 338, 339, 398, 451, 452, 460, 499, 500 browsing 212, 220, 221, 272 see also grazing Brown bear (Ursus arctos) 142, 143, 144, 146, 634 Bulgaria 11, 108, 111, 143, 275, 443, 444, 550, 609 Butzer, Karl 336, 337

Index Büyük Menderes River, Turkey 18, 263, 388 Byzantine Period 272, 503 see also Early Byzantine Tectonic Paroxysm C3 93, 98, 127 C4 93, 98, 127 Cairo 74, 389, 416, 418 Calabria 7, 10, 11, 12, 13, 97, 113, 119, 141, 171, 179, 190, 390, 398, 448, 484, 495, 496, 497, 508, 514, 535 Calabrian arc 5, 7, 9, 12, 13, 16, 385, 386, 387, 390, 391, 398, 500 see also subduction; tectonics and landscape development Calabrian pine 206, 207, 210, 220 calcium carbonate 53, 93, 98, 100, 115, 124, 127, 128, 171, 178, 185, 211, 258, 263, 264, 266, 267, 268, 287, 288, 290, 292, 293, 296, 297, 298, 299, 301, 305–6, 308, 312, 313, 315, 333, 393, 394, 423, 424, 442, 445, 476, 586 see also calcrete; limestones; travertine; tufa; speleothems calcrete 24, 127, 178, 244, 305, 333 see also cemented till; cementation Calderone glacier 80, 357, 362 California, USA 69, 157, 223, 440, 541, 546, 547, 553, 554, 615, 627, 628 Camargue wetlands, France 213, 263, 274, 621, 623, 629, 642 Campanian volcanic province 436, 442–6 see also volcanoes and volcanism Canary Islands 148, 154, 155, 207, 630 Cantabrian Mountains, Spain 108, 123, 143, 353, 354, 359 see also Picos de Europa carbon and the carbon cycle 91, 93, 127, 197, 416, 599, 608 see also carbon dioxide inorganic carbon 197, 599, 603, 608, 609, 611 organic carbon 53, 93, 127, 197, 567, 574 carbon dioxide 80, 91, 93, 107, 112, 197, 217, 222, 223–4, 290, 297, 303, 437, 439, 440, 442, 449, 599, 603, 644 Carboniferous limestones 295 Carpathian Mountains 7, 11, 13, 141 Catalonia 19, 85, 97, 109, 190, 217, 248, 423, 514, 516, 517, 521, 522, 525, 526, 528, 530, 531, 532, 533, 535, 536, 541, 546, 547, 549, 552, 556, 596 catastrophic flooding 74, 122, 194, 323, 442, 515, 518, 519, 520, 522, 523,

525, 526, 530, 531, 535, 537 see also storms and floods Caucasus Mountains 10, 108, 149, 255, 298, 378, 435, 436 Caves and cave deposits 56, 102, 103, 104, 123, 141, 146, 154, 156, 157, 174, 249, 287, 288, 290, 293, 295, 296, 297, 298–300, 301, 305, 307, 309, 310–13, 314, 315, 357, 390, 393, 394, 395, 402, 403, 415, 421, 422, 443, 444, 451, 484, 571, 633, 635–6 see also rockshelters; speleothems; karst Cave salamander (Proteus anguinus) 157–8, 636 Cave bear (Ursus spelaeus) 146 cemented till 365, 368, 370 see also cementation; calcrete cementation, 179, 292, 298, 303, 305, 306, 308, 321, 322, 325, 332, 333, 334, 393 see also calcrete; cemented till Cenozoic Era 79, 91, 92, 93, 94, 98, 104, 255, 295, 307, 324, 445 see also global cooling cereals 240, 618, 633 Chad, Lake 55, 418 chamois (Rupicapra rupicapra) 640 channel incision 7, 15–6, 18, 22–6, 195, 295, 301–2, 306, 319, 321–2, 327–8, 330, 333–5, 338–9, 345–6, 367, 389, 481, 483, 529 see also river systems and environmental change; gully development channel management see river systems and environmental change charcoal 102, 112, 206, 208, 223, 270, 312, 332, 544, 546, 553, 617 check dams 239–244 see also dams; reservoirs chemical weathering see weathering and rock breakdown; dissolution Chile 69, 157, 541, 615 chotts 261, 262, 405, 416–7, 421, 424 chronosequence 174 circumpolar vortex 119 see also jet stream cities see urban environments and urbanization Classical Period 272, 306, 319, 335, 403, 435, 448, 449, 503, 564, 565 see also Roman Period; Bronze Age; Ancient Greece clays and clay minerals 19, 20, 51, 54, 172, 173, 174, 178, 185, 189, 272, 297, 402, 416, 420, 422, 425, 439 clearance, woodland 158, 192, 208, 271, 272, 336, 563, 572, 574, 575, 577, 618 see also deforestation

653

climate: air temperature 69–71, 74, 77–81, 83–4 cyclones and cyclogenesis see storms and floods future climate 80–5 heatwaves 73, 77–9 hurricanes (‘Medicanes’) and tornadoes 532–5 modelling 37, 47, 54, 55–6, 58,83–6, 91, 94, 99, 103, 169, 176, 223, 232–3, 515, 519–23, 542, 554, 555, 576–7, 603, 605–6, 610–1 orographic influence on 35–6, 57, 93, 103, 515, 521–2, 532, 534, 603, 605 rainfall 19, 39, 54, 69–75, 79–80, 82–86, 99–102, 120–3, 126–127, 171, 179–180, 183–191, 204, 212, 220, 229–245, 265, 290, 295, 303, 309, 344, 361, 425, 439, 445, 513–37, 548, 552, 555, 568–71, 574, 576, 583–5, 592, 602, 610–11, 643 seasonality 37, 39, 45–7, 55, 71, 73–5, 80, 85, 89, 91, 93, 97, 99–102, 117, 126–7, 178, 204, 229, 232–6, 247, 249–50, 264, 267, 276, 295, 417, 523, 532–3, 546, 548 snowfall 69–70, 91, 150, 174–5, 190, 268, 296, 301, 303, 354–78, 439, 527, 534–5 see also snowmelt; snowline winds 38–9, 40, 42–44, 55, 69–70, 74–5, 77–8, 93, 119, 179, 386, 388, 415–25, 457, 513–5, 521, 531–35 see also storms and floods; Bora; Etesian; Levante; Mistral; Tramontana climate change see climate; Quaternary environmental change; Heinrich Events; Dansgaard-Oeschger events; Milankovitch cycles; Little Ice Age; Medieval Warm Period coastal environments 385–407 see also sea-level change aquifers 307, 309, 346 beaches 386, 393, 398, 402, 405–6 biology 395–7 caves 102, 312, 394, 395, 403 coastal flooding 404, 406, 527 currents 40–4, 386, 388, 389 deltas 15–16, 18, 40, 83, 95, 151, 214, 263, 272, 274, 277, 319, 336, 346, 347, 386, 388–9, 393, 402–6 dunes 386, 405, 422, 423–4 harbours 77, 394, 395, 398–400, 402–4, 405, 471, 477, 480, 482, 493, 498, 501

654

Index

coastal environments (cont.) Holocene development 401–2 see also deltas littoral cells 386–88, 389 plains 77, 97, 109, 197, 319, 325–7, 338, 344, 390, rocky shores 23, 222, 287, 288, 289, 315, 386, 395, 396 tectonics 23, 36, 385–93, 398–400 see also tsunamis wetlands 83, 95, 97, 213, 255, 257–60, 263, 272–3, 275, 278, 279, 281 winds 77, 78 coins 321, 476 see also artefacts colluvial processes see hillslopes Como Lake, Italy 255, 258, 262, 273, 278 coniferous forest 95–7, 117–8, 127, 141, 203–5, 207–8 see also forest communities conservation 126, 139, 145, 160, 206, 212, 213, 225, 237, 239, 245, 255, 260, 267, 272–4, 276, 279–80, 281, 312–4, 563, 564, 565, 571, 572, 578–9, 583, 592, 615, 621, 624, 627, 632, 637, 640–3, 645 see also biodiversity and conservation Constantinople 504 see also Istanbul Corinth, Gulf of, Greece 7, 21, 22, 23, 390, 400, 401, 471, 479, 480–1, 484, 485, 495, 496, 502–3, 505, 506, 507, 508, 510 Corinth, Greece 20, 470, 486 Corsica 10, 11, 40, 158, 206, 232, 323, 324, 329, 344–5, 353, 354, 370, 373, 416, 417, 420, 421, 424, 516, 572, 624, 626, 628, 631, 632–3, 635, 641, 642 Cosquer cave, France, 141, 394, 395, 633 Côte d’Azur, France 358, 400, 496–7, 505, 508, 516–7, 535 crayfish 257, 272–3, 278–9, 628 Cretaceous 11, 16, 93, 127, 196, 293, 307 Crete 21, 35, 39, 113, 151, 154, 175, 206, 207–8, 212–5, 216, 220, 221, 224, 249, 263, 293, 323, 328, 330, 338, 340, 343, 344, 345, 354, 355, 397, 399, 400, 420, 421, 422, 424, 441, 450, 454, 460, 471, 479, 480–3, 485, 487, 493, 495, 496, 497–501, 505, 530, 547, 572, 575, 600, 603, 607, 608, 609, 616, 631, 632, 644 Croatia 11, 77, 143, 169, 261, 274, 275, 288, 293, 296, 298, 301, 306, 312, 314, 353, 369, 385, 386, 532, 535, 543, 587, 588, 594,

631, 636, 640, 642 see also Dalmatia cryosphere 91, 92, 319, 353–78 see also glacial and periglacial environments cryptic refugia, see refugia cultivation see agriculture cultivars 160, 638–9 see also agriculture; domestication cyclones and cyclogenesis see storms and floods; air masses Cyprus 10, 11, 33, 39, 41, 42, 70, 75, 154, 169, 207, 215, 233, 236, 260, 263, 275, 385, 386, 391, 400, 401, 418, 424, 485, 495, 496, 500, 504, 505, 508, 515, 534, 537, 587, 588, 594, 595, 596, 616, 621, 631, 634, 638, 640, 644 Cyrenaica 215, 255, 324, 325–327, 616, 630, 644 see also Libya; Tripolitania Dalmatia 78, 143, 288, 309, 422, 498, 531, 636 see also Croatia dams (natural and artificial) 20, 39, 40, 86, 97, 193, 194, 232, 239–44, 247, 249, 251, 257–60, 272, 274, 276, 279, 281, 314, 337, 345, 346, 367, 369, 405, 471, 481, 483, 485, 525, 527, 530, 583, 585, 589, 590, 591, 592 see also reservoirs; check dams Dansgaard-Oeschger events 45, 57–8, 331 see also sub-Milankovitch climate variability dating methods see lichenometry; luminescence; orbital tuning; potassium-argon; radiocarbon; uranium-series Dead Sea 56, 79, 102, 103–4, 105, 119, 255, 257, 261, 265, 266, 268, 272, 276, 321, 473, 474, 476, 505 see also Lisan, Lake debris flows 18, 20–1, 193, 437, 439, 445, 500, 505, 535 see also mass movements deciduous forest 92, 95–8, 100–101, 107, 109–10, 115–7, 127, 151, 173, 203–6, 216, 222, 271–2, 556, 572, 623–4, 626 deep water 34, 35, 37, 40, 42–4, 52–3, 94, 98, 99, 119, 122 see also Mediterranean Sea; North Atlantic Deep Water (NADW) deflation 175, 262, 305, 415, 423–5, 571 see also aeolian processes and landforms deforestation 80, 237, 239, 270, 272, 273, 274, 281, 314, 415, 425, 522, 536, 571–2, 575, 591, 615, 618, 632, 643 see also clearance

deglaciation 268, 375, 392 see also glacial and periglacial environments dehesas 158, 206, 225, 617 deltas see coastal environments desalination 70, 587 deserts see aridity; aeolian processes and landforms desertification 204, 220, 249, 415, 564, 565–6, 579, 615, 618, 643 see also aridification diatoms 123, 267, 268, 270, 273, 279 Dinaric Alps 39, 149, 255, 290, 314, 353, 354, 357, 368, 531–2, 583 Dinaric karst 157, 296, 301, 303, 304 dissolution 37, 171, 175, 179, 185–6, 258–9, 261, 264, 287, 290–2, 305, 314, 396, 416 see also weathering; solute loads DNA 125, 139, 142, 216 domestication 154, 157, 160, 220, 563, 638, 639, 642 drought 37, 69, 72, 73, 77, 79, 83–86, 100, 103, 113, 117, 123, 126, 128, 173, 204, 209, 218, 223, 249, 259, 265, 269, 319, 531, 535, 537, 548, 550, 555, 564, 568, 572, 592, 600, 610, 628, 631, 632 see also climate drylands see aridity dunes see aeolian processes and landforms; coastal environments dwarfism in mammals 154, 157, 634 Early Byzantine Tectonic Paroxysm 400, 471, 477 see also earthquakes and seismicity earthquakes and seismicity 5, 7, 11–3, 20–3, 190, 325, 390, 399–400, 401, 437, 444, 446, 450, 452, 457, 459, 469–87 see also earthquakes and tsunamis; tectonics and landscape development in Antiquity 471, 474–6, 481–4 data sources and methods 474–6 hazard management 485–7 impacts 478–84 earthquakes and tsunamis 493–510 see also coastal environments Eastern Mediterranean Deep Water (EMDW) see Mediterranean Sea Ebro delta, Spain 18, 388, 402, 404, 405 Ebro River basin, Spain 15, 16, 20, 39, 151, 186, 250, 262, 265, 329, 360, 389, 402, 416, 425, 516, 527, 530, 569, 585 eccentricity see Milankovitch cycles ecosystems 83, 90, 150, 151, 157, 159, 203–225, 234, 237, 249–50, 267, 269–70, 274–81, 287, 314, 378, 415, 439, 541, 543, 546, 550–53, 567–9, 575, 578, 586, 590, 599,

Index 615–8, 621, 624–8, 638, 643, 645, 646 Egypt 10, 57, 58, 74, 77, 78, 84, 145, 213, 258, 269, 272, 274, 275, 277, 287, 288, 293, 295, 301, 305, 307, 386, 389, 400, 401, 415, 418, 422, 424, 425, 471, 477, 480, 483, 485, 508, 583, 586, 587, 588, 590, 593, 594, 595, 596, 631, 632, 640 El Niño 73, 574 endemism 125, 140, 143, 146, 148, 149, 150, 151, 154–9, 203, 211–2, 215–6, 266–7, 274, 279, 282, 615, 622, 623, 626, 628–36, 641, 643–5 see also biodiversity Eocene 91–7, 98, 144, 291, 300 ephemeral streams and lakes 17, 232, 234, 244–5, 247–8, 250, 258, 261, 263, 277, 305, 325, 421, 422, 523, 525, 527, 585 see also river systems and environmental change Ephesus, Turkey 338 Epirus, Greece 19, 109, 151, 240, 321, 322, 367, 640, 641 Equilibrium Line Altitude (ELA) 354, 358, 362, 363–4, 366, 369–70, 373 see also snowline; glacial and periglacial environments erosion see soils; hillslopes; incision essential oils 204, 218–20 Etesian wind 74, 77, 78 see also climate; storms and floods Etna, Sicily 13, 14, 17, 172, 180, 335, 354, 357, 435, 436, 438, 441, 449–50, 458, 459, 460 see also volcanoes and volcanism Euboea, Greece 338, 397 European Commission (EC) 313, 544, 550, European Union (EU) 86, 224, 241, 250, 267, 280, 487, 535, 536, 564, 575, 599, 605, 609, 635, 636, 637, 641 Eurasia and Eurasian plate 5, 10–11, 14, 36, 45, 94, 139, 144–6, 147, 255, 293, 324, 435, 469, 472, 474, 563 eutrophication 257, 259, 260, 266, 274, 276, 277, 280–1, 313, 591, 599, 635 see also lakes and wetlands Euphrates River see Tigris-Euphrates evaporation 5, 14, 33, 37, 39–43, 45, 48, 52, 55–6, 78, 83, 92, 100, 102–3, 119, 124, 218, 219, 230, 231, 239, 247, 263, 265, 267–268, 276, 303, 305, 389–90, 519, 567, 569, 585, 590, 592, 602, 610–1 see also climate evaporites 7, 14–5, 36, 94, 127, 171, 250, 261, 264, 276, 292, 389, 425 see also Messinian Salinity Crisis

evapotranspiration 220, 232, 233, 314, 552, 555, 569, 571, 584, 643 see also evaporation; transpiration Evros River, Greece 151 Evrotas River, Greece 328, 330 see also frontispiece extensional tectonics 8–11, 13, 17, 21–3, 36, 195, 385, 451, 473, 480 see also tectonics and landscape development extinction 90, 95, 97, 111–3, 125, 139–40, 142, 145–7, 154–5, 157, 276, 621–2, 626–7, 629–37, 639–40, 642 extinct volcanoes 450 Faraya-Mzaar Mountains, Lebanon 70 farming see agriculture faulting and faults 13, 15, 16, 17, 18, 19, 20, 21, 22, 23, 36, 189, 196, 256, 261, 290, 291, 293, 294, 295, 299, 325, 372, 388, 390, 402, 446, 469, 470, 472, 473, 474, 476, 477, 478, 479–80, 481, 482, 486, 487, 493, 494, 497, 498, 502, 504–505 see also tectonics and landscape development Ferdinandea Island 10 fire see wildfires fish fauna and habitats 143, 154–5, 158–9, 267, 272, 279, 281, 347, 394, 402–3, 621, 627–8, 635–6 see also fisheries and fishing fisheries and fishing 83, 257–60, 266, 272–3, 276–9, 281, 406, 485, 501, 503 flash flooding see storms and floods flint tools 312 see also stone tools flooding see storms and floods flood hazard see storms and floods flood management see storms and floods floodplains 234–5, 242, 245, 246–9, 263, 322, 338, 344–7, 393, 402 see also river systems and environmental change; gravel bed rivers; channel incision; sediment loads and yields fluvial terraces see river terraces; channel incision flysch 171, 246, 294, 298, 300, 325, 333, 404, 455 Fontaine de Vaucluse 16, 288, 296, 298 see also karst foraminifera 47, 50, 51, 54, 56, 92, 103, 122, 123 see also oxygen isotope records forest clearance see deforestation forest communities (past and present) 90, 93–8, 107, 114–7, 123–5, 141–2, 146–8, 150–1, 155–8, 172–3, 177, 180, 183, 187, 203–9, 216, 218,

655

222–4, 237, 246, 249–50, 269–74, 329, 336, 439, 532–3, 541–556, 563–4, 569–72, 574–5, 577, 617–27, 629, 632, 640–4 see also coniferous forest; deciduous forest; rainforest forest fires see wildfires fossil fuels 603, 609 France 16, 37, 74, 77, 78, 84, 86, 109, 114, 116, 124, 141, 146, 159, 160, 171–4, 179–81, 183, 185, 187, 192, 194–5, 206, 211, 213, 217, 222–4, 232–3, 236–7, 249, 255, 263, 273–5, 287–8, 293, 295, 296, 298, 301, 307, 312, 314, 339, 344, 358, 360, 374, 386, 394, 398, 417, 418, 421, 424, 436, 459, 471, 485, 487, 513–4, 516, 521–2, 525, 527, 530, 532–7, 543–4, 552, 554, 563, 566, 583, 587–8, 594–6, 600, 621–6, 628, 631–5, 640–2 see also Corsica Franchthi cave, Greece 288, 312, 402–3, 444, 451, 484 frost and frost action 120, 173, 178, 292, 310, 353, 357, 360–1, 378 see also permafrost; weathering and rock breakdown fuel (for burning and wildfires) 93, 157, 218–9, 220, 543, 546, 548–50, 552–6 see also fossil fuels, ignition FYR Macedonia 10, 11, 233, 256, 257, 258, 274, 275, 279, 367, 479, 587, 594 Galicia, Spain 169, 301, 375 Garda Lake, Italy 257, 262, 280 garrigue 181, 203, 206, 208, 212, 223, 233, 563 gastropods 272, 393, 396, 500 see also molluscs Genoa, Gulf of 38, 39, 74, 75, 77, 78, 80, 390, 497, 515, 537 genetic diversity 109, 112, 113, 117, 125, 126, 142, 143, 152, 160, 216, 623, 624, 628, 638–40, 645 see also genetic divergence genetic divergence 90, 125, 143 see also speciation geoarchaeology 319, 322, 393, 394–5 geochronology see dating methods geophytes 212, 214, 218, 626, 632 geothermal energy 305, 435, 440, 442 Gibraltar 74, 288, 294, 324, 391, 631 Gibraltar arc 11, 12, 13, 15 Gibraltar, Straits of 7, 14, 15, 33, 34, 35, 36, 40–3, 47–48, 77, 78, 94, 119, 144, 211, 386, 389, 390, 401, 406, 532, 601, 610 GISP2 57, 105, 121 see also GRIP

656

Index

glacial and periglacial environments 353–78 Holocene and Little Ice Age records 361–64 interaction with other geomorphological systems 377–8 see also river systems and environmental change modern environments and processes 354–61 Pleistocene records 364–77 Global Positioning Systems (GPS) 8, 10, 11, 195, 472, 473, 476, 480 global cooling 57, 91–2, 95, 119, 122–3, 174, 197, 331, 361 see also Quaternary environmental change global warming and impacts 78–80, 83, 92, 119, 268, 398, 406, 564, 566, 576–7, 599, 610, 624, 632, 643, 644 see also climate; greenhouse gases goats 146, 154, 157, 220–1, 290, 336, 563, 570, 575, 637, 639–40, 642 goethite 176 see also soils gorges 141, 171–2, 194–5, 212–3, 292, 300–302, 305–6, 309–11, 315, 323, 332, 335, 340, 343, 389, 638, 641 Granada, Spain 109, 177, 193, 416, 422, 572 grasslands 93, 107, 116, 146, 151, 203–5, 208–9, 212, 216–7, 223, 550, 556, 570, 578, 617, 626, 631, 634, 643 see also steppe gravel bed rivers 234–5, 240, 242, 245, 297, 298, 305–6, 308, 325, 327, 332, 334, 345, 402 see also river systems and environmental change; sediment loads and yields grazing 184, 203, 204, 212, 214, 216, 217, 218, 220, 224, 225, 237, 239, 249, 272, 273, 438, 546, 548, 549, 553, 554, 563, 567, 571-2, 573, 575-7, 579, 619, 631, 644 see also browsing; herbivory; overgrazing; pastoralism gravel extraction 235, 345–7 Great auk (Pinguinus impennis) 141, 633 Greece 7, 10–11, 19–20, 21–3, 36, 70, 75, 77, 80, 84, 95, 98, 100, 102–5, 107, 109, 111, 113–4, 116–8, 120–2, 123, 124, 125, 127, 143, 146, 150–1, 155, 171, 173, 174, 178,179, 185, 188, 206–7, 212, 213, 215–6, 230, 232, 233, 235, 239, 240, 245, 249, 255, 257–9, 261–2, 266–7, 274, 275, 279–280, 288, 290–1, 296, 303, 306, 311–2,

313, 315, 319–24, 328, 330, 331–3, 337–9, 344, 346–7, 353, 355, 359, 362, 364, 365–72, 376–8, 385, 386, 399–400, 402–3, 405–6, 422, 424, 442–3, 450–7, 459, 469–72, 474, 477, 478–83, 484, 485–7, 497–504, 506–7, 514, 522–3, 530, 533, 534, 541, 543–4, 554, 563, 566, 569, 572, 575, 587, 588, 594, 595, 596, 599–600, 608, 616, 621, 628, 630, 631, 632, 635, 640–1, 642, 644 see also Crete greenhouse effect 592, 599 see also climate; greenhouse gases greenhouse gases 80, 83, 89–90, 92, 529, 599, 607, 609 see also global warming and impacts Greenland ice cores, see GISP2 and GRIP GRIP 57, 329, 330 see also GISP2 groundwater 70, 79, 194, 230–3, 249, 251, 255, 263, 267, 272, 274, 276, 280, 282, 290, 295–8, 301, 303, 305, 307, 309, 312–4, 394, 406, 424, 439, 446, 527, 564, 571, 583, 585–7, 589–94, 621 see also aquifers; water resources Guadalentin River, Spain 243, 249 Guadalfeo River, Spain 193, 249 Guadalope River, Spain 328–30 see also Ebro River basin gully development 5, 18, 20, 22, 23–5, 185–6, 189–90, 246–7, 337, 570, 573, 575 gypsum 16, 20, 26, 171, 185–6, 261, 263–4, 272, 292–3, 297, 305, 423, 425 see also evaporites habitats: aquatic 54, 56, 157, 266, 276, 280–1, 313, 404, 396, 622–43 diversity 149–151, 153, 157, 159, 160, 209, 212, 214–7, 287, 617, 643–6 human impact 146, 158, 217, 274, 276, 313, 346–7, 554, 616–43 terrestrial 107, 120, 125, 139, 141, 145, 147, 149, 156, 209, 618, 622–43 Hadley Cells 38, 122, 126–7, 602 Hadley Centre Regional Climate Model, HadRM3 83–5 haematite 175–6 see also iron oxides halophytes 97 haplotypes 111, 143 harbours see coastal environments hazards see specific hazards heathlands 203–4, 211, 547, 575 heatwaves see climate Heinrich Events 45, 57, 119, 268, 312, 330–1

Hellenic arc 7–9, 12–3, 21, 255, 385–7, 390–1, 393, 473–4, 479, 496–500, 506–10 see also subduction Hellenides, Greece 7, 10, 108, 293 herbaceous vegetation 93, 97–8, 107, 116, 127, 205, 216–7, 219, 548, 626, 633 herbivory 204, 217–8, 556, 575, 633, 637 Hérault River, France 172 Herculaneum see Pompeii Herzegovina see Bosnia hillslope-channel coupling 324, 344 hillslopes 169, 179–97 see also mass movements hillslope terracing see agricultural terraces Holocene 12, 13, 20, 25, 47–9, 55–6, 57–8, 79, 89–90, 99–100, 102–3, 115–7, 122–4, 139, 141–2, 147, 157–8, 178–9, 208, 213, 217, 222–4, 263, 268–72, 303, 307, 310, 312, 315, 320–3, 335–44, 353, 361–4, 366, 369, 378, 388, 390–3, 397–400, 410, 403–4, 423, 441–2, 446, 456, 473, 481–2, 499, 529, 541, 546–8, 564, 575, 617, 634, 638 hydrology see climate; hillslopes; runoff; river regimes; storms and floods Iberian lynx (Lynx pardellus) 156, 642 Iberian Mountains 532, 359–60, 364–5, 375 Iberian Peninsula 11, 21, 36, 37, 71, 74, 75, 98, 102, 111, 120, 139, 141, 142, 144, 158, 159, 206, 212, 222, 225, 233, 239, 254, 261, 324, 329, 331, 359–60, 364–5, 375, 376, 515, 521, 546, 547, 568, 584, 596, 617 see also Spain; Portugal Ibex (Capra ibex) 145, 156, 633–4 ice sheets 37, 47, 90–2, 95, 98, 102, 104, 107, 113, 122, 125, 262, 331, 373–4, 376, 392 402, 435 see also glacial and periglacial environments; ice caps ice caps 47, 299, 354–5, 364, 367, 375 ice cores see GISP2, GRIP, Antarctic ice cores ice rafting see Heinrich Events ignition 218, 220, 543, 546–9, 553–4, 556, 609 see also fuel; wildfires incision see channel incision; gully development; river terraces India 73, 91, 602–4 Indian monsoon 39, 46, 55, 56, 73, 74–5, 602–3 Indian Ocean 7, 14, 94, 509–10, 603

Index interglacials 7, 45, 46, 47, 51, 55, 56, 57, 58, 90, 92, 94, 97–100, 101, 102, 104–7, 114–9, 122–5, 125, 127, 140, 141, 203, 212, 215, 216, 268, 312, 319, 320, 321, 325, 327, 334, 377, 390, 391, 546 see also Holocene; Quaternary environmental change; Last Interglacial insolation 45–6, 54, 55, 56, 58, 98, 99, 100, 101, 102, 106, 115, 116, 117, 118, 119, 123, 125, 127, 175, 222 see also climate; Milankovitch cycles Inter-governmental Panel on Climate Change (IPCC) 80, 83, 527 interrill erosion 183, 185 Intertropical Convergence Zone (ITCZ) 38, 99, 175, 176, 178, 602, 603, 605 invasion and invasive species 146, 147, 157, 207, 209, 214, 220, 222, 616, 626–8, 629, 630, 633, 644 invertebrates 148–149 Ioannina Lake basin, Greece 90, 100, 101, 103, 104, 105, 107–8, 109, 115, 117, 120, 121, 124, 125, 216, 259, 261, 262, 362, 365, 369, 377, 572 Ionian Islands 401, 471, 477, 485 Ionian plate 21, 22 Ionian Sea 10, 15, 22, 41, 42, 44, 74, 76, 77, 78, 107, 385, 387, 388, 400, 404, 406, 471, 472, 474, 479, 480, 486, 505, 508 Iran 255, 257, 259, 261, 274, 486, 602 Iraq 75, 86, 255 Iron Age 158, 223, 271, 339 iron oxides 175, 176, 178 see also terra rossa; soils irrigation 40, 70, 79, 83, 232, 234, 239–40, 249, 250, 257, 260, 269, 272, 274, 276, 277, 278, 280–1, 345, 347, 564, 571, 584, 585, 589–90, 591, 592, 596, 635 Israel 18, 19, 56, 74, 80, 97, 99, 103, 104, 116, 119, 171, 173, 175, 176, 178, 182, 184, 206, 207, 208, 209, 215, 217, 220, 232, 233, 236, 244, 247, 255, 257–60, 274, 275, 276, 277, 280, 288, 312, 313, 315, 386, 389, 394, 405, 418, 421, 422, 423, 424, 471, 477, 480, 496, 498, 500, 504, 505, 535, 541, 563, 587, 588, 589, 593, 594, 595, 596, 616, 621, 625, 631, 636, 640, 642, 644 Istanbul 278, 279, 478, 504 see also Constantinople Italy 7, 8, 10, 11, 16–7, 19, 20, 21, 36, 38, 69, 75, 77–80, 83, 85–6, 92, 97,

102, 109, 113, 116–117, 119–20, 123–4, 141, 157, 171, 173, 177–8, 185, 188–9, 192–3, 195–6, 206, 222, 232–4, 236, 255, 257–62, 267–9, 270, 274–8, 280, 287–8, 292, 295, 298, 303, 312–4, 319, 331, 339, 345, 357–8, 375–6, 385–6, 391, 395, 400, 406, 421–3, 435–7, 440–1, 442–50, 457, 459, 469–71, 474, 477, 478, 480, 484, 487, 495–7, 506–10, 513–4, 521–2, 525, 527, 529–30, 532–7, 541, 543–4, 554, 565–6, 569, 572, 587–8, 594–6, 600, 621–2, 630–1, 635, 640, 642 see also Appenines, Calabria; Sicily; Sardinia ITCZ see Intertropical Convergence Zone jet stream 38, 73, 122, 531, 537, 602–4 Jerusalem 312, 420, 535 Jordan 206, 275, 631, 642 Jordan River 236, 265, 276, 277, 280, 483, 485 Jordan Valley and rift 56, 255, 256, 257, 261, 265, 266, 268, 272, 276, 295, 321, 324, 418, 469, 471, 480, 504, 505 Julian Alps 331, 354, 357, 362, 378 Jura Mountains 11 Jurassic Period 94, 127, 293, 294, 295 karst 7, 16, 26, 287–315 see also tufa; travertines; speleothems caves 298–300, desert karst 303, 305 early history of karst research 290 fauna 156, 157 glacio-karst 291, 314, 353, 364, 367, 377 see also glacial and periglacial environments habitats 150 hydrology 232, 295–8, lakes 257–60, 261, 262, 264, 266, 272, 273, landforms 290–310 landscapes 255, 290–5, 300–307 records of environmental change 310–13 threats and conservation 312–4 vadose and phreatic systems 295–300 weathering 171, 174, 292 Kastritsa cave, Greece 103, 104, 105, 108 katabatic air 37, 77, 78 see also climate Konya basin, Turkey 258, 260, 261, 268, 274, 275, 416, 423, 621 Kopais basin, Greece 90, 105, 109, 120, 121 Köppen classification of climate 69 see also climate

657

Korana River, Croatia 288, 306 Krka, River, Croatia 306 Lago Mare 14, 36 see also Mediterranean Sea lagoons 23, 40, 83, 153, 255, 258, 260, 263, 272–3, 278, 279, 280, 293, 386, 393, 402, 406 see also lakes and wetlands lake levels 47, 56, 57, 102–5, 119, 123, 268–70, 278, 280, 340, 342 lakes and wetlands 255–82 biology 266–7 hydro-chemistry 263–66 long records of change 267–72 origin and morphology 256–63 recent trends 272–4 threats and conservation 274–81 lake sediments 23, 116, 178, 262, 272, 280–281, 420, 444, 546, 563 laminated sediments 37, 48, 51, 102, 123, 258, 267–8, 447 land abandonment 212, 214, 217, 224, 271, 485, 549, 553, 555, 566, 571, 572–5, 618 land degradation 563–79 see also habitats; soils; biodiversity and conservation contemporary perceptions 565–7 history 563–4 processes and precursors 567–77 prospects for improvement 577–9 landslides, 20–21, 24–6, 73, 79, 86, 189–90, 192–6, 334, 378, 437, 439, 441, 445, 448–9, 471, 482–5, 493, 497, 501–3, 505, 510, 527, 534–5, 565 see also mass movements Languedoc-Roussillon, France 516, 522, 525, 536 La Niña 73 Last Glacial Maximum (LGM) 45, 47–9, 55, 102–4, 106, 107–12, 114, 268, 375–6, 393 Late-glacial 20, 102, 116, 127, 141, 171, 268, 321, 327, 333, 367, 370, 373, 375 Last Interglacial 7, 99, 100, 101, 115–7, 122, 123–4, 325, 327, 390, 391, 546 lava 175, 257, 335, 355, 437, 438–42, 444, 446–51, 453–5, 458–9 see also volcanoes and volcanism; magma and magmatism Lebanon 70, 108, 149, 207, 209, 215, 255, 275, 295, 299, 364, 386, 418, 480, 495, 504, 505, 542, 587, 594, 595, 616, 631, 632, 640, 644 Lesvos, Greece 174–5, 216, 575 Levant 40, 78, 99, 145–6, 169, 232–3, 239, 255, 261, 269, 271, 287, 293,

658

Index

295, 326, 385, 386, 392–3, 400, 406, 415, 418, 424, 495, 504–5, 563–4, 583 Levante, Spain 78, 516, 521, 532, 536–7 Levante wind 77, 78, 532, 537 see also climate Levantine Sea 33, 34, 36, 39, 41, 42, 44, 385–7, 388, 496, 504–5, 508, 510 levées 20, 454 Libya 10, 55, 58, 78, 145, 156, 169, 215, 232, 241, 274, 275, 287, 288, 295, 301, 305, 315, 319, 324, 325–9, 330–1, 386, 416–8, 422, 424, 470, 477, 480, 485, 508, 583, 586–9, 592–6, 630, 631, 640 see also; Cyrenaica; Tripolitania lichens 173, 244, 422 lichenometry 340, 343, 344, 362, 364 Liguria 496–7, 508, 514, 516, 521, 525, 536 Ligurian Alps 108, 215, 616, 644 Ligurian Sea 38, 39, 75, 387 limestone 7, 16, 25–6, 171, 174–5, 178, 179, 233, 255, 261, 287–315, 321, 325, 327, 332–3, 353, 365, 367, 378, 385, 394, 396, 399, 416, 450, 455, 476, 586, 626, 636, 641 see also karst; calcium carbonate Lions, Gulf of 35, 37, 39–40, 42–3, 103, 197, 401, 405, 424 Lisan, Lake 103, 105, 119, 257, 268 see also Dead Sea Lisbon 69, 71, 480 Little Ice Age (LIA) 79, 249, 272, 321, 340–44, 348, 353, 357, 361–4, 404 littoral zone 250, 274, 276, 307, 329, 331, 385, 386–90, 392–407, 415, 439, 523, 530 see also coastal environments Llobregat River, Spain 516, 523–7, 530, 535 long profiles of river channels 325, 335 see also river systems and environmental change; tectonics and landscape development Lost Eden hypothesis 214–5, 566, 615 luminescence dating 321, 323, 329, 422 optically stimulated luminescence dating (OSL) 325, 328, 330, 334 thermoluminescence (TL) dating 332, 375 Lyell, Charles 398, 450 Macedonia, Greece 10, 151, 276, 338, 472, 477, 485, 621 Maggiore Lake, Italy 255, 257, 262, 273, 278, 280

Maghreb 10, 11, 16, 107, 169, 233, 239, 245–7, 255, 325, 331, 421, 565, 566, 571, 583, 588, 619, 630 magma and magmatism 10, 13, 14, 436, 439, 441–51, 452, 454–8 see also volcanoes and volcanism Majorca 240, 299, 309, 310, 554, 586 see also Balearic Islands Malta 10, 14, 70, 275, 386, 401, 533, 587–9, 594, 596, 621, 631, 634, 638, 640 Mammal fauna 93, 141, 143–4, 145–7, 154–8, 214, 218, 626, 628–9, 633–4, 638, 642 mammoths 145, 146, 633 maquis 151, 173, 188, 203–4, 206–7, 209, 211, 212, 214, 233–4, 337 marine environment see Mediterranean Sea marine gateways 5, 7, 14–5, 91, 94, 389 see also Mediterranean Sea marine terraces 5, 7, 22, 390 see also coastal environments Maritime Alps 108, 215, 354, 358, 362, 373, 375–6, 378, 616, 621, 626, 644 marls 7, 19, 26, 173, 178, 180, 183, 185, 189, 246, 248, 250, 263, 272, 293, 325, 333 Marmara, Sea of 274, 390, 495–6, 504, 505, 506, 508, 510 see also Black Sea Marseille 16, 74, 77, 141, 395, 398, 633 marshes 23, 151, 213, 255–6, 260–1, 263, 272–3, 274, 276–7, 280–1, 386, 402, 536, 590, 621, 623, 628, 636, 642 see also lakes and wetlands mass movements 5, 7, 17–8, 20–1, 26, 169, 188–91, 197, 311, 514 see also debris flows; landslides; pyroclastic density currents and deposits Massif Central 77, 109, 113–5, 436, 532 mattoral 147–8, 150, 152, 156, 180, 183–4, 190, 203, 206, 249, 618, 625, 627, 629 MEDALUS 249, 577 Medieval Period 322, 335, 338, 339, 546 see also Medieval Warm Period Medieval Warm Period 79, 321, 341 Mediterranean Sea: present and past 33–58 basin topography and water circulation 33–6 centennial- to millennial-scale variability 57–8 evolution of the basin 36–7 glacial cycles 47–8

modern climate and oceanography 37–40 monsoon maxima and sapropels 48–57 Quaternary change 44–58 water circulation and deep water formation 40–44 Megara basin, Greece 22–3 Mesolithic Period 158, 310, 314, 338, 339, 401, 403, 406, 451 Mesozoic Era 10, 33, 36, 94, 255, 261, 295, 307, 324, 445, 455 Messina, Italy 471, 478, 497 Messina, Straits of 400, 495–6, 497, 505, 508–10 Messinian Salinity Crisis 5, 7, 14–6, 19, 23, 26, 36, 94–5, 127, 144, 151, 171, 195, 261, 287, 292, 296, 307, 388–90 see also tectonics and landscape development; evaporites metal mining 345–6 metal pollution 274, 277, 280, 282, 345–6, 439, 442, 450, 450–1, 638 Methana, Greece 13, 450–1 methane 80, 124, 406, 599, 603, 607 see also greenhouse gases microclimate 108, 140, 171, 173, 179, 212, 215, 217, 357 microplates 139, 144, 293, 295, 324 see also tectonics and landscape development Middle East 39, 74, 75, 77, 83, 84, 86, 146, 147, 204, 324, 418, 420, 422, 471, 478, 608, 610, 630, 640 Mid-Pleistocene Transition (MPT) 37, 106, 114 migration (flora and fauna) 90, 94, 95, 107, 109, 112, 113, 120, 140, 142, 144–6, 151, 152, 154, 155, 156, 212, 214, 234, 266, 274, 276, 281, 624, 635, 637–8 see also invasion and invasive species Milankovitch cycles 45, 58, 89, 90, 124, see also sub-Milankovitch climate variability eccentricity 37, 45–6, 54, 91, 92, 98–9 obliquity 37, 45–6, 91–2, 98–9, 104, 113, 117, 119 precession 45–6, 54, 97, 98–102, 104, 117–9, 125, 127 Minoan Civilization 214, 436, 441, 443, 452, 454, 460, 493, 499–501, 505 Miocene 5, 10, 11, 13, 14, 21, 92, 93, 94, 95, 97, 98, 126, 127, 139, 142, 144, 145, 178, 216, 256, 261, 305, 307, 455 see also Messinian Salinity Crisis

Index Mistral 35, 37, 40, 43, 75, 77, 78, 532 see also climate; storms and floods molluscs 266, 267, 396, 402, 403 see also gastropods monsoon 36, 37, 39, 40, 45, 46, 48–58, 73, 74, 75, 93, 98, 99, 100, 102, 104, 116, 602, 603, 604 see also sapropels; climate Montenegro 257, 263, 264, 266, 275, 288, 289, 290, 291, 292, 301, 306, 308, 314, 353, 354, 356, 357, 358, 367, 373, 374, 378, 495, 497, 508, 587, 588, 594 Monticchio, Lago Grande di 90, 109, 115, 120, 261, 268, 270, 572 moraines 329, 353, 357, 362–76, 377, 378 sea also glacial and periglacial environments; till Morocco 10, 11, 14, 15, 19, 21, 36, 39, 42, 43, 58, 94, 108, 109, 116, 123, 150, 176, 205, 206, 212, 232, 233, 234, 236, 241, 246, 247, 248, 259, 261, 269, 272, 273, 275, 287, 288, 303, 328, 330, 354, 360, 386, 416, 417, 424, 480, 536, 541, 542, 543, 564, 565, 587, 590, 591, 594, 595, 596, 618, 620, 621, 631, 632, 636, 640, 642, 643 see also Atlas Mountains mudflows 189, 249, 437, 439 see also mass movements Murcia, Spain 178, 189, 240, 242, 243, 245, 249, 250, 514, 530, 573, 576 Nearctic region 142, Near East 17, 78, 80, 89, 100, 104, 107, 116, 147, 209, 213, 222, 223, 224, 319, 364–5, 423, 566, 583, 584, 588, 593, 596, 618, 622, 632 Negev Desert 19, 247, 248, 416, 422, 423, 425 Neogene 23, 24, 49, 91, 127, 145, 293, 481 Neolithic Period 146, 147, 154, 158, 218, 223, 315, 322, 335, 338, 339, 398, 402, 403, 442, 456, 563 Nestos River, Greece, 151 Nile River 15, 16, 17, 18, 36, 39, 40, 54, 55, 58, 83, 97, 99, 100, 116, 176, 230, 232, 233, 258, 263, 272, 277, 386, 388, 389, 402, 405, 420, 422, 530, 584, 585, 588, 590, 594, 596 nitrous oxides 80, 599, 600, 605, 607 see also air quality nivation 174, 302, 353, 357, 357, 364, 365, 378 see also glacial and periglacial environments North Africa 10, 11, 33, 37, 39, 55, 58, 74, 75, 77, 78, 84, 89, 99, 104,

107, 116, 142, 144, 145, 146, 147, 148, 149, 150, 159, 204, 208, 209, 223, 224, 232, 239, 247, 255, 261, 262, 263, 276, 281, 293, 303, 305, 306, 310, 315, 319, 324, 325–9, 331, 385, 386, 392, 400, 401, 415, 416, 418–25, 496, 508, 531, 532, 536, 543, 548, 549, 553, 554, 564, 566, 583, 600, 604, 610, 618, 621, 630, 634, 636, 640, 642, 643 see also Algeria; Egypt; Libya; Maghreb; Morocco; Tunisia North Atlantic Deep Water (NADW) 35, 94, 119, 122 North Atlantic Ocean 7, 8, 12, 14–5, 23, 33, 35, 36–8, 39, 40–2, 45, 47, 57–8, 69–70, 73, 74, 80, 91, 94, 98, 99, 104, 119–20, 122–5, 141, 147, 230, 268, 269, 312, 329, 331, 340, 341, 342, 346, 389, 390, 406, 424, 469, 472, 509, 515, 518, 521, 534, 601, 602, 605, 608, 610 see also Heinrich Events; North Atlantic Deep Water; North Atlantic Oscillation North Atlantic Oscillation, NAO 38, 73, 80, 340, 415, 574, 601, 608, 610 Northern Hemisphere 37, 38, 45, 46, 57, 58, 69, 70, 91, 92, 96, 122, 142, 146, 605, 633 oases 97, 288, 305, 571 obliquity see Milankovitch cycles obsidian 447, 451, 456 Older Fill 321–3 see also river systems and environmental change Oligocene 8, 10, 11, 16, 21, 91–3, 95–8, 126, 145, 216 olive 77, 151, 160, 184, 206–7, 210, 214–6, 220, 271–2, 298, 336–7, 572, 633, 639 Olympia, Greece 319–21, 338, 346, 483 Olympus, Mount, Greece 108, 354–5, 366–7, 372 ophiolite 333, 353 orbital forcing see Milankovitch cycles orbital tuning 46, 94, 369, 436 see also Milankovitch cycles; dating methods orogeny 5–6, 10–1, 197, 255, 331, 334, 473 see also tectonics and landscape development orographic influence on climate see climate; storms and floods Ottoman Period 271, 638 outwash sediments 372 see also glacial and periglacial environments overbank flows 247, 527, 529 see also storms and floods; floodplains overbank sediments 247–8, 333, 340–1 see also floodplains

659

overgrazing 158, 208, 272, 405, 564, 570, 571, 577, 615, 618, 632 see also grazing; pastoralism overkill hypothesis 146, 634 overland flow 182–3, 185, 239, 244, 247, 249, 575 see also runoff; hillslopes oxygen isotope records: ice cores 57, 120, 328, 330, marine 36, 46, 47, 53, 54, 55, 56, 58, 91, 92, 95, 98, 105, 106, 113, 114, 118, 120, 121, 122, 124, 268, 312, 329, 365, 369, 376, 390 lake 123, 124, 265, 268, 269, 272 speleothems 312, 313 ozone 122, 217, 599, 604–6, 607, 608 see also air quality; TOMS Pamukkale, Turkey 288, 306, 308, 314, 315 see also travertine; tufa Palaearctic 141–2, 146, 147, 149, 155 palaeofloods and palaeohydrology 272, 339–45 see also river systems and environmental change Palaeolithic Period 157, 310–1, 314, 319–321, 333, 347, 385, 394–5, 401–3, 406, 443–4, 563, 633 Palaeolithic art 314, 395, 633 palaeoshorelines 266, 267–78, 390–403 see also sea-level change; lake levels; marine terraces palaeosols 93, 98, 127, 174, 177–8, 421–2, 424 see also soils Palestine 160, 232, 255, 423, 496, 500, 505, 564, 587 palynology see pollen and pollen records Panama, Isthmus of 91 panarchy 565, 573, 577–8 Pantelleria 10, 13, 263, 435, 436, 449–50 see also volcanoes and volcanism paraglacial 378 see also glacial and periglacial environments parent material see soils pastoralism 179, 405, 423, 564, 575, 577, 616–7, 625 see also grazing passerines 152, 624 see also bird fauna Passer Valley, Italy 193, 195 peat 116, 280, 393, 443, 571 pedogenesis see soils Peloponnese, Greece 239, 321, 323, 337–8, 403, 474–5, 498, 503, 544, 632, 640 see also Corinth; Sparta Pergusa Lake, Sicily 259, 263–4, 268 periglacial processes see glacial and periglacial environments perihelion 45, 46, 54 permafrost 357–60, 378 see also glacial and periglacial environments photosynthesis 93, 204, 224, 552

660

Index

Piave River, Italy 194, 345 Picos de Europa, Spain 288, 295, 298–9, 303, 359, 364, 375, 378 see also Cantabrian Mountains Pindus Mountains, Greece 19, 21, 69, 103, 107–8, 113, 120, 155, 212, 301, 303, 314, 331, 353–4, 355, 357, 364–7, 369, 377–8 Piva River (Sinjac), Montenegro 298, 367 plate tectonics see tectonics and landscape development playa lakes 258, 260, 261–3, 425 see also saline lakes and wetlands Pleistocene see Mid-Pleistocene Transition (MPT); Quaternary environmental change; Last Glacial Maximum (LGM) Pliny the Younger 435, 444, 632 Plinian and subplinian eruptions 435, 442, 444–6 Pliocene 10, 13, 14–16, 24, 36, 49, 91–5, 97–8, 106, 112, 114, 116–8, 125, 139, 142, 145–7, 177, 305, 315, 333, 335, 404, 451, 546 ploughing 182, 222, 239, 249, 336 see also tillage Po River, Italy 39, 44, 389 Po delta 18, 151, 263, 388, 404, 406 pollen and pollen records 47, 56–7, 89–128, 174, 177, 206, 209, 211, 214, 216, 222, 268–72, 328–30, 339, 365, 376–7, 546, 571–2 pollution, see air quality; water quality Pompeii and Herculaneum 444–6 see also volcanoes and volcanism Pontic Mountains, Turkey, 108, 149, 354, 355, 364, 378 Portugal 124, 178, 180, 185, 206, 211, 216–7, 219, 223, 233, 275, 280, 360, 375, 541, 543–5, 554, 566–7, 599–600, 617, 628, 631, 640 Portugese margin 120–1, 124 potassium-argon dating 335, 355 pottery 321, 456, 476 see also artefacts Precambrian 324, 325 precession see Milankovitch cycles precipitation see climate progradation see coastal environments; sea-level change Provence 249, 393, 470, 471, 485, 516, 517, 535 Provençal basin 7, 9, 10, 15 Pyrenees 10, 19, 70, 79, 80, 84, 108, 109, 141, 144, 149, 171, 178, 212, 293, 298, 301, 314, 325, 331, 354, 358–9, 360, 362–4, 373–5, 376, 377, 378, 404, 417, 515, 516, 521, 522, 523, 524, 532, 573

pyroclastic density currents and deposits 437, 438, 441, 442, 444, 445, 446, 447, 452, 455, 456, 457, 458, 500, 505 Qattara depression, Egypt 416, 425 Quaternary environmental change: see also Last Glacial Maximum (LGM); Heinrich Events; Dansgaard-Oeschger events; interglacials; Last Interglacial aeolian processes 415–25 climate and vegetation change 98–128 coastal environments 388–402 glacial and periglacial environments 361–77 history of vertebrate fauna 139–160 karst dynamics 310–13 marine environments 44–58 river systems and environmental change 325–40 seismicity 469–84 tectonics and landscape development 5, 7, 15, 16–26, volcanism 436, 441–57 radiocarbon dating 104, 141, 312, 321–3, 332, 335, 338, 340–2, 362, 369,374,375–6,400,444,454,505 AMS dating 104, 323, 332, 340–2 databases 340–2 raindrop impact 179–80, 183–4, 219, 572 see also soils rainfall see climate rainforest 94, 95 raised beaches 23, 481 see also marine terraces; palaeoshorelines sea-level change ramblas 24, 26, 186, 242–3, 245, 249 see also wadis Ramsar convention and designated sites 256–60, 274–5, 277, 640 raptors 152, 637, 642 see also bird fauna reafforestation 237, 271 see also afforestation recharge see groundwater red beds 321–2 see also terra rossa Red deer (Cervus elaphus) 145, 642 Red Sea 10, 33, 47, 56, 99–100, 144, 261, 390, 418, 472 refugia 90, 107–14, 125–6, 140–3, 212, 216, 621, 643, 645–6 relictual taxa 113 Renaissance 313 reservoirs see also dams; water resources evaporation losses 590 sedimentation 193, 239–42, 244, 247, 345, 571, 591–2 water storage 79, 194, 196, 239–43, 277, 279, 345, 585, 589–92, 642

Rhodes 18, 33, 42, 471, 473, 478, 482, 495, 498–9 Rhône River and valley, France 17, 35, 37, 39, 77, 86, 151, 230, 233, 236, 344, 404, 527, 584, 585, 586, 621 Rhône River delta and sub-marine fan 18, 20, 263, 346, 388, 393, 402, 404, 405, Rhône River palaeovalley 15, 16, 195, 389 RICAMARE 644 Rifian corridor 7, 14, 15 see also Rif Mountains; Betic-Rif region Rif Mountains 10, 11, 19, 24, 150, 215, 616, 644 see also Betic-Rif region rifting 10, 195, 261, 324, 385, 436, 472 see also rift valleys rift valleys 261, see also rifting; Jordan Valley and rift rillenkarren 292, 310 see also karst; weathering and rock breakdown rills and rill erosion 185–186, 187, 239, 249 see also hillslopes riparian habitats 95, 250, 346, 585, 621, 627, 628, 629, 632 see also wetlands river capture 7, 16, 21, 23, 24–6, 333–4, 335 river channels 150, 189, 235, 248, 319, 321, 325, 336, 337, 340, 344–7, 480, 481, 483, 522, 523, 529, 537, 584 river regimes 47, 55, 229–34, 235, 237, 239, 240, 243, 248, 250, 315, 321, 323, 344, 523, 536, 584–6 see also climate river systems and environmental change 319–48 see also sediment loads and yields Holocene records 335–340 Little Ice Age 349–344 models of change 321–4 Pleistocene records 325–35 see also river capture; river terraces recent human impacts 344–7 tectonic setting for 324–35 see also river capture river terraces 15, 18, 321, 322, 325, 327, 328, 329, 330, 333, 334, 335, 336, 344 rock breakdown see bedrock erosion and weathering rockshelters 108, 287, 297, 310, 321 see also rockshelter sediments, caves and cave deposits rockshelter sediments 174, 297, 298, 310–2, 313, 333, 378, 484 Roman Period 160, 179, 239, 240, 247, 263, 271, 272, 276, 281, 305, 306, 321, 322, 327, 335, 336, 339, 395,

Index 398, 399, 400, 404, 405, 442, 445, 446, 451, 476, 480, 485, 529, 546, 563, 564, 617, 628, 632 see also Classical Period Roman triad 158, 616 Romania 11, 143, 275, 550, 609 Roman volcanic province 436, 442 see also volcanoes and volcanism Rome 69, 71, 236, 276, 306, 404, 437, 442, 621 see also Roman Period runoff 26, 39, 54, 55, 56, 57, 99, 102, 104, 151, 158, 178, 180, 182–7, 214, 229, 232, 233, 235, 237, 239, 240, 244–5, 249, 251, 272, 280, 314, 319, 331, 337, 477, 536, 552, 553, 566, 569, 570, 572, 573, 574, 575, 579, 583, 584, 587, 591, 618, 626, 643 see also hillslopes; climate; storms and floods; river regimes; Mediterranean Sea Sahara Desert 55, 58, 69, 75, 77, 97, 99, 116, 141–2, 145, 152, 255, 263, 360, 415–22, 424–5, 531, 571, 577, 583–4, 586, 592, 637 Saharan dust see aeolian processes and landforms Sahel 55, 58, 74,116, 602, 610 saline lakes and wetlands 94, 123, 213, 250, 258, 260–7, 269, 274, 276–8, 280, 346, 389, 394, 425, 485, 571, 592 saline soils 211, 249 see also salinization salinization 70, 179, 272, 274, 276, 346, 571 salt deposits see evaporites salt weathering 425 Santorini (Thera) 13, 269–70, 400, 435–6, 441, 443, 451–6, 459–60, 485, 493, 495–6, 499–502, 505 sapropels 33, 36, 46, 47, 48–57, 98–102, 104, 116, 123, 128, 267, 269, 312–3 see also Mediterranean Sea Sardinia 7, 10, 42, 232, 263, 324, 329, 393, 394, 417, 420–1, 424, 514, 516, 521, 547, 571, 631 satellite imagery and data 76, 108, 406, 417, 476, 534, 544–5, 550, 553, 555, 565, 576, 600–1, 605, 608–9 see also TOMS savanna and savanna-like habitats 58, 93, 99, 109, 116, 146, 149, 203–4, 206, 246, 577, 615 sclerophyllous vegetation 90, 97–8, 100–1, 109, 115–7, 126–8, 173, 203–7, 212, 222–4, 249, 546, 623

scree 171–2, 178, 205, 212–3, 303, 321, 322, 324, 370, 378, 641, 643 see also talus sea-level change 7, 14–6, 18, 21–3, 36, 47–8, 80, 83, 91–2, 94, 103, 141, 195, 213, 263, 303, 305, 307, 312, 323, 338, 385, 388, 390–406, 423, 425, 482–3, 497, 531, 643 sebhkas 261–2, 405, 424–5 sediment budgets 248, 298, 300, 346, 347, 386 sediment supply 54, 235, 319, 323, 331–3, 335, 340, 344, 353, 378, 402, 483 see also hillslope-channel coupling; sediment budgets sediment loads and yields see also reservoirs suspended sediment 17, 23, 26, 40, 54, 58, 186, 195, 214, 237, 242, 245–8, 297, 420, 591 bed load sediment 17, 23, 26, 195, 247–8, 297, 569–71 Segre River, Spain 516, 530 seismicity 5, 7, 10, 11–13, 15–6, 18, 20–23, 26, 139, 189, 324–5, 329, 331, 387, 399–401, 404, 439, 441–2, 444–52, 454–5, 457, 459, 469–87, 493–4, 497, 500–2, 504–9 see also earthquakes; tsunamis Serbia 11, 45, 367, 587–8, 594 Sicily 10, 11, 12, 13, 17, 20, 77, 97, 113, 117, 144, 169, 206, 222, 247, 259, 263, 264, 268, 325, 335, 336, 357, 390, 395, 401, 415, 418, 435, 446, 449–50, 472, 476, 495, 496, 497, 498, 508, 514, 571, 631, 632, 634 see also Etna Sicily–Cap Bon line 139 Sicily, Strait of 10, 33, 34, 35, 41, 42, 52, 75, 401 Sierra Nevada, Spain 108, 174, 212, 249, 255, 354, 359–60, 364, 378 Sinai 149, 324, 326, 386, 416, 418, 422–3, 425 sinkholes 303, 305, 484 see also karst Sirocco wind 74, 77, 415, 418, 513, 531, 537 see also climate; storms and floods Sirte basin 10, 388, slackwater deposits 311, 333, 340–1, 348 see also palaeofloods slope deposits see scree; hillslopes slope failure see mass movements Slovenia 11, 109, 143, 169, 293, 298, 314, 331, 357, 362, 377, 525, 533, 541, 587–8, 594, 636, 640 snowline 354–5, 357, 364, 375–6 see also glacial and periglacial environments, Equilibrium Line Altitude (ELA)

661

snowmelt 91, 192, 439, 527 soils see also terra rosse aggregates 173, 185 carbon 172, chronosequence 174 conservation 245, 565, 578–9 erosion 179–188, 190–192, 197, 209, 214, 219, 237, 247, 249, 271–2, 336–7, 552–3, 563, 566, 574, 576–7, 579, 618, 626, 638, 643 forming processes 173–4, 176–7 moisture 102, 180, 206, 217, 232, 240, 247, 583 nutrients 173, 566 organic matter, 172–3, 220, 301, 566, 572, 574 parent material 169, 171, 174–175, 178 profiles 172, 176, 178–9, 290, 332 resources 178 solifluction 179, 355, 357, 360, 376 see also glacial and periglacial environments; hillslopes solute loads 176, 246, 263, 265, 290, 297, 301, 305, 442 see also dissolution; sediment loads and yields Sorbas basin, Spain 14, 24–6, 171, 245, 288, 293, 333–4 Soreq Cave, Israel 56, 90, 102–3, 123, 288, 312–3 see also speleothems South Africa 69, 157, 222, 541, 546–7, 552, 575, 578, 615, 627, 628 Spain 5, 7, 10–11, 13, 18–20, 23–26, 41, 70, 75, 77–9, 80, 84, 86, 94, 108–9, 116, 123, 146, 156, 169, 171, 173, 174, 177–8, 180, 182, 183, 185–93, 205, 206, 209, 211–13, 217, 232, 237, 240–5, 248–50, 258–65, 266, 267, 269–70, 274–6, 277–79, 280, 288, 293, 298–9, 301, 309, 314, 319, 324, 328–31, 333–5, 339–42, 344–6, 353, 359–60, 362–4, 373–6, 388, 390, 401, 405–6, 415–6, 420–1, 423–5, 436, 477, 513–4, 516–7, 519, 521–2, 528–38, 541, 543–44, 546–9, 552, 554–5, 564, 566–9, 572, 574, 576, 586, 587, 590, 594, 605, 623, 625, 631, 640 Sparta, Greece 471, 485, 502 see also frontispiece speciation 90, 125–6, 139–40, 148, 160, 216, 624 see also allopatric speciation; genetic divergence species diversity 139, 143, 217, 220, 396, 615–6 see also biodiversity

662

Index

speleothems 56, 101–2, 103, 119, 123, 175, 297, 298, 299, 305, 312–3, 315, 319, 393–4, 420 steppe 37, 93, 97, 107, 115, 118, 122, 141, 147, 203, 208–10, 212, 223, 246, 268, 329–30, 564, 618–9, 634, 643 stone-walled terraces see agricultural terraces stone tools 311, 319, 321, 444 see also Palaeolithic Period storms and floods 513–38 see also catastrophic flooding classification of floods and storms 523–27 climate variability and flooding 527–31 convective systems and storms 518–23 cyclones and cyclogenesis 38, 39, 58, 71, 74–77, 126, 513, 515, 517, 518–9, 521, 522, 531, 532, 534, 535, 536, 537, 600, 605 see also anticyclones hailstorm events 534 hazard mitigation 536–7 heavy snowfall 534–5 hurricanes 534 impact of extreme events 535–6 tornadoes 532–4 wind storms 531–2 strontium isotopes 102, 420 subduction 5, 7–17, 21, 36, 293–4, 385, 387, 389–91, 435, 450, 456, 474, 478, 498 see also back-arc basins; Calabrian arc; Hellenic arc; tectonics and landscape development sub-Milankovitch climate variability 45, 58 see also Dansgaard-Oeschger events; Heinrich Events; Little Ice Age; Medieval Warm Period sub-tropical high pressure 37, 38–9, 69, 73, 122, 126–7, 514 sub-tropical jet 73, 122, 537, 602, 604 succession, ecological 101, 112, 115–20, 127, 151, 156, 217, 223, 549, 556, 565, 578, 626 Super-Sauze earthflow, French Alps 191, 195 suspended sediment see sediment loads and yields Syria 75, 86, 109, 116, 142, 146, 209, 215, 255, 259, 271, 275, 298, 386, 418, 471, 483, 495, 505, 587, 594–6, 616, 631, 634, 636, 640, 642, 644 Tabernas basin, Spain 18–9, 187 talus 302, 309, 359, 366, 378 see also scree Tara River, Montenegro 367

Taurus Mountains, Turkey, 149, 212, 255, 301, 354–5, 364–5, 378 tectonics and landscape development 5–32 see also earthquakes and seismicity; volcanoes and volcanism; coastal environments global setting and history 5–7 , 14 see also Messinian Salinity Crisis geodynamics of the Mediterranean 7–14 see also earthquakes and seismicity; volcanoes and volcanies temperature see climate, climate change Tenaghi Philippon, Greece 107, 109, 113–5, 118–9, 120–1, 174, 261 tephra 128, 269–70, 311–2, 425, 436–9, 441–8, 450–2, 454–8, 499–500 see also volcanoes and volcanism Ter River, Spain 516, 528, 530 terraces see agricultural terraces; marine terraces; river terraces terra rossa 175, 178, 298, 305, 309, 315, 325, 327, 415–6 see also soils Tertiary Period 10, 20, 94–8, 112–3, 116, 125, 126–8, 139, 142, 178, 212, 255, 305, 425, 631 Tethys 5, 7–10, 14, 15, 36, 94, 95, 97, 127, 144, 145, 171, 255, 261, 293, 294, 385, 386, 387, 472 see also marine gateways Thera see Santorini Thermaikos Gulf, Greece 347 thunderstorms see climate; storms and floods Tibetan high 602, 603 Tibetan Plateau 37, 91, 197 Tigalmamine Lake, Morocco 109, 116, 123, 259, 261, 269, 272, 273 Tigris-Euphrates River 56, 86 till see glacial and periglacial environments; moraines tillage 182–3, 214 see also ploughing; soils TOMS (Total Ozone Mapping Spectrometer) 417–8, 419, 420 see also aeolian processes and landforms; aerosols Tortonian 11, 14, 15, 16, 19, 92, 94 tourism and recreation 70, 79, 83, 257, 258, 272, 273, 274, 276, 279, 280, 313, 314, 385, 404, 405, 406, 423, 425, 437, 438, 446, 448, 449, 457, 458, 459, 485, 523, 534, 554, 592, 593, 618, 621, 622, 632, 643 Tramontana wind 78, 531–2, 537 see also climate; storms and floods transgression, marine 23, 399, 401, 402, 403 see also sea-level change

transhumance 212, 563, 575, 637–8 see also pastoralism; grazing transpiration 204 see also evapotranspiration travertine 258, 297, 298, 305–6, 308, 315 see also tufa; speleothems Trieste, Italy 78, 288, 290, 301 Tripoli, Libya 288, 301, 401, 416, 422 Tripolitania, Libya 325, 327, 415 Troodos Mountains, Cyprus 70 Troy, Turkey 338, 485, 496, 504 tsunamis 20, 400–1, 437, 441, 446, 448–9, 455, 460, 470–1, 480, 482, 483–4, 493–510 quantification 493–4 major tsunami events in the Mediterranean 400–1, 494–505 tsunami generation 505 hazard assessment 505–8 risk mitigation 508–9 tufa 97, 261, 297, 298, 299, 302, 305–6, 308, 315, 321, 322 see also calcium carbonate; calcrete; travertine; speleothems Tunis 401, 416 Tunisia 10, 19, 33, 58, 75, 109, 116, 144, 150, 176, 213, 217, 232, 239, 248, 260, 274, 275, 276, 292, 315, 328, 330, 339, 360, 386, 393, 400, 401, 405, 415, 417, 418, 421, 423, 424, 425, 508, 515, 516, 530, 533, 534, 563, 571, 585, 586, 587, 589, 590, 591, 592, 594, 595, 596, 621, 631, 640, 642 turbidity and turbidity currents 20, 40, 346, 396, 460, 496, 586 see also turbidites turbidites and mega-turbidites 20, 400 Turkey 8, 10, 11, 18, 21, 39, 74, 75, 77, 80, 83, 84, 86, 102, 127, 159, 169, 171, 174, 176, 207, 212, 222, 232, 233, 234, 236, 255, 256, 257, 259, 260, 261, 262, 263, 266, 267, 268, 270, 271, 274, 275, 276, 278, 279, 287, 288, 293, 295, 298, 301, 305, 306, 308, 314, 319, 325, 331, 335, 354–5, 362, 364–5, 376, 385, 386, 390, 400, 405, 406, 418, 423, 435, 441, 454, 455, 469, 470, 471, 472, 477, 478, 480, 481, 486, 487, 499, 500, 504, 515, 533, 534, 535, 550, 571, 587, 588, 594, 595, 596, 615, 630, 631, 632, 640, 642, 643 Tuscany 13, 172, 182, 436, 442, 496, 497, 508, 516, 521, 621 Tyrrhenian Sea and coast 10, 13, 15, 36, 43, 102, 122, 215, 312, 388, 395,

Index 400, 448, 478, 496, 497, 506, 508, 515, 516, 521 Tyrrhenian basin 8, 9, 12, 13, 42, 95, 385, 387, 509 UN (United Nations) 565, 566 uplift (crustal) 7, 14, 15, 16, 17–8, 19, 21–3, 24, 36, 37, 77, 91, 103, 173, 189, 195, 197, 255, 261, 290, 292, 293, 302, 305, 325, 329, 333, 335, 389, 392, 397, 399, 401, 444, 446, 454, 456, 470, 473, 476, 477, 479, 480, 481, 482, 483, 498 see also tectonics and landscape development uranium-series (U/Th) dating 24, 103, 113, 303, 305, 306, 312, 315, 321, 323, 325, 328, 330, 331, 332, 333, 365, 368, 370, 376 urban environments and urbanization 79, 86, 240, 246, 260, 274, 276, 278, 279, 281, 313, 314, 345, 347, 404, 405, 424, 437, 449, 459, 485, 522, 523, 525, 529, 530, 535, 536, 555, 564, 586, 591, 596, 605, 609, 622, 625, 637, 643, 644 USA 18, 553 Vaiont dam disaster 179, 194–6 see also mass movements Valencia 249, 280, 423, 530, 641 Valencia basin and trough 7, 9, 10, 21, 385, 387, 401 valley fill sediments see river systems and environmental change valley floor environments 193, 235, 242, 302, 309, 321–3, 325, 327, 329, 331, 334–5, 337–8, 346–8, 641 see also deltas; floodplains; riparian habitats Van, Lake, Turkey 102, 123, 255, 257, 261, 263, 267–70, 281 van Andel, Tjeerd 337 varves 119, 257–60, 267–9, 272 see also laminated sediments vegetation and ecosystem dynamics 203–225, 567–70 vegetation history 89–128, 208, 213, 217, 222–4 see also pollen and pollen records Venice 75, 77, 83, 406, 531 Vera basin, Spain 24 Vesuvius 13, 435, 436, 437, 438, 441, 442, 444–6, 457–60 see also volcanoes and volcanism Vikos–Aoos National Park, Greece 640, 641 Vikos gorge, Greece 108, 212, 230, 235, 288, 296, 300, 302, 306, 323, 324, 640, 641

vines 183, 190, 214, 272, 572, 618 Voidomatis River, Greece 235, 296, 300, 323–4, 328, 330, 331–3, 377 volcanoes and volcanism 5, 7, 12–4, 17, 20, 22, 128, 139, 171–2, 319, 435–460 ash see tephra gas and aerosols 439–41 hazards 180, 269, 436–441 lake basins 89, 255, 258–63, 267 lava flows 438–9 provinces and geological setting 435–6, 441–57 pyroclastic density currents 438–9 risk management 457–460 tephra clouds and falls 438 tsunamis 20, 437, 441, 446, 448–9, 455, 460, 493, 497, 499–501, 505–6, 509 volcanism and: archaeological records 442–445 climate 79, 90, 128 coasts 398–400 mountain glaciation 355, 364 river behaviour 17, 335 waders 151–3, 637 see also bird fauna wadis 55, 99, 234, 325, 236, 327–9, 330, 405 see also river systems and environmental change warblers 147, 148–50, 152, 276, 618, 625 see also bird fauna Water Framework Directive (WFD) 250–51, 280, 321, 564, 579, 596 see also European Union water harvesting 239 see also agricultural terraces water resources 84, 86, 229, 235, 241, 247, 279, 280, 281, 299, 309, 313, 315, 345, 552, 563, 564, 565, 583–96 see also runoff competition and threats 590–1 geography 583–6 human impact on 586 management and sustainability 591–2 resource by country 586–8 resource by population 588–9 threat of climate change and temporal trends 592–6 water table 194, 211, 280, 295, 296, 297, 301, 304, 309, 312, 314, 346, 425, 442, 642 see also groundwater water quality 246, 250, 251, 274, 280, 281, 290, 296, 313, 314, 315, 437, 564, 586, 591, 592, 594 see also eutrophication wave action 77, 305, 385, 386, 394, 395, 396, 441, 450, 455, 460, 483,

663

484, 493–510, 532, 535 see also tsunamis; marine terraces; coastal environments wave cut platform see marine terraces weathering and rock breakdown 23, 102, 169–78, 197, 287, 292, 298, 300, 303, 305, 306, 307, 309, 310, 311, 321, 333, 366, 370, 416, 425 see also bedrock erosion and weathering; dissolution westerlies 36, 37, 102, 126, 602 see also climate Western Mediterranean Deep Water (WMDW) 34, 35, 40, 42, 43 see also Mediterranean Sea wetlands 110, 150, 151, 213, 232, 246, 255–267, 272–82, 345, 346, 402, 406, 571, 590, 603, 618, 621, 622, 627, 628, 629, 631, 640, 642, 643 see also lakes and wetlands wheat 160, 272, 639 White Mountains, Crete 216, 354 White Desert, Egypt, 305, 307 Wild boar (Sus scrofa) 146, 640 wildfires 541–56 controls on fire regime 548–9 extent of the problem 541–3 hazard and impact on ecosystems 550–53 hazard mitigation 553–5 historical context and data sources 546–8 magnitude and frequency 543–6 winds 78 see also climate, Bora; Etesian; Mistral; Sirocco; Tramontana wind storms 513, 515, 531–7 see also storms and floods Wolf (Canis lupus) 146, 154, 635, 640 Würm glaciation 141, 146, 362, 369–70, 373–5, 376 see also glacial and periglacial environments xerophytes 97, 576, 625–6 see also halophytes xerosols 176, 177, 178 see also soils Younger Fill 321–3, 335–336 see also river systems and environmental change Yugoslavia 11, 77, 85, 232, 272, 279, 298, 301, 314, 331, 355, 362, 365, 367, 386, 479, 541, 543, 638 Zagros Mountains 10, 12, 149, 255 Zewana, Wadi, Libya 325, 326–30 Zohary’s law 209, 222