DEVELOPMENTS IN EARTH & ENVIRONMENTAL SCIENCES
4
MEDITERRANEAN CLIMATE VARIABILITY
VOLUME 1
GEOSCIENCES, ENVIRONMENT AND MAN by H. Chamley
VOLUME 2
MEDICAL GEOLOGY EFFECTS OF GEOLOGICAL ENVIRONMENTS ON HUMAN HEALTH by M.M. Komatina
VOLUME 3
OIL POLLUTION AND ITS ENVIRONMENTAL IMPACT IN THE ARABIAN GULF REGION by M. Al-Azab, W. El-Shorbagy and S. Al-Ghais
Developments in Earth & Environmental Sciences, 4
MEDITERRANEAN CLIMATE VARIABILITY Edited by
P. LIONELLO University of Lecce, Italy
P. MALANOTTE-RIZZOLI MIT, Cambridge, MA 02319, USA
R. BOSCOLO Clivar, International Project Office, Southampton, UK
Amsterdam – Boston – Heidelberg – London – New York – Oxford – Paris San Diego – San Francisco – Singapore – Sydney – Tokyo
ELSEVIER B.V. Radarweg 29 P.O. Box 211, 1000 AE Amsterdam, The Netherlands
ELSEVIER Inc. 525 B Street, Suite 1900 San Diego, CA 92101- 4495 USA
ELSEVIER Ltd. The Boulevard, Langford Lane Kidlington, Oxford OX5 1GB UK
ELSEVIER Ltd. 84 Theobalds Road London WC1X 8RR UK
ß 2006 Elsevier B.V. All rights reserved. This work is protected under copyright by Elsevier B.V., and the following terms and conditions apply to its use: Photocopying Single photocopies of single chapters may be made for personal use as allowed by national copyright laws. Permission of the Publisher and payment of a fee is required for all other photocopying, including multiple or systematic copying, copying for advertising or promotional purposes, resale, and all forms of document delivery. Special rates are available for educational institutions that wish to make photocopies for non-pro¢t educational classroom use. Permissions may be sought directly from Elsevier’s Rights Department in Oxford, UK: phone (þ44) 1865 843830, fax (þ44) 1865 853333, e-mail:
[email protected]. Requests may also be completed on-line via the Elsevier homepage (http://www.elsevier.com/locate/permissions). In the USA, users may clear permissions and make payments through the Copyright Clearance Center, Inc., 222 Rosewood Drive, Danvers, MA 01923, USA; phone: (þ1) (978) 7508400, fax: (þ1) (978) 7504744, and in the UK through the Copyright Licensing Agency Rapid Clearance Service (CLARCS), 90 Tottenham Court Road, London W1P 0LP, UK; phone: (þ44) 20 7631 5555; fax: (þ44) 20 7631 5500. Other countries may have a local reprographic rights agency for payments. Derivative Works Tables of contents may be reproduced for internal circulation, but permission of the Publisher is required for external resale or distribution of such material. Permission of the Publisher is required for all other derivative works, including compilations and translations. Electronic Storage or Usage Permission of the Publisher is required to store or use electronically any material contained in this work, including any chapter or part of a chapter. Except as outlined above, no part of this work may be reproduced, stored in a retrieval system or transmitted in any form or by any means, electronic, mechanical, photocopying, recording or otherwise, without prior written permission of the Publisher. Address permissions requests to: Elsevier’s Rights Department, at the fax and e-mail addresses noted above. Notice No responsibility is assumed by the Publisher for any injury and/or damage to persons or property as a matter of products liability, negligence or otherwise, or from any use or operation of any methods, products, instructions or ideas contained in the material herein. Because of rapid advances in the medical sciences, in particular, independent veri¢cation of diagnoses and drug dosages should be made.
First edition 2006 Library of Congress Cataloguing in Publication Data A catalogue record is available from the Library of Congress. British Library Cataloguing in Publication Data Mediterranean Climate Variability. - (Developments in earth & environmental sciences ; 4) 1. Mediterranean climate 2. Climatic changes - Mediterranean Region I. Lionello, P. II. Malanotte-Rizzoli, P. III. Boscolo, R. 551.6’91822 ISBN-10: 0 - 444 -52170 - 4 (this volume) ISBN-13: 978 - 0 - 444 -52170 -5 (this volume) ISSN: 1571-9197 (series)
The paper used in this publication meets the requirements of ANSI/NISO Z39.48 -1992 (Permanence of Paper). Printed in The Netherlands.
Abridged Contents
Introduction. The Mediterranean Climate: An Overview of the Main Characteristics and Issues
1
1.
Mediterranean Climate Variability over the Last Centuries: A Review
2.
Relations between Climate Variability in the Mediterranean Region and the Tropics: ENSO, South Asian and African Monsoons, Hurricanes and Saharan Dust
149
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
179
Changes in the Oceanography of the Mediterranean Sea and their Link to Climate Variability
227
5.
The Atlantic and Mediterranean Sea as Connected Systems
283
6.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
325
7.
Regional Atmospheric, Marine Processes and Climate Modelling
373
8.
The Mediterranean Climate Change under Global Warming
399
Subject Index
417
3.
4.
27
This page intentionally left blank
Contents
Abridged Contents Preface Introduction. The Mediterranean Climate: An Overview of the Main Characteristics and Issues 1. The Mediterranean Region: Climate and Characteristics 2. Regional Processes and Links to the Global Climate 3. Climate Trends and Change at Regional Scale 4. Socio-environmental Aspects Acknowledgements References
v xiii
1 1 6 11 14 17 18
Chapter 1. Mediterranean Climate Variability over the Last Centuries: A Review 27 1.1. Introduction 29 1.2. Past Regional Mediterranean Climate Evidence and Extremes 32 1.2.1. Evidence from Early Instrumental and Documentary Data 32 1.2.2. Evidence from Natural Proxies 57 1.3. Large-Scale Climate Reconstructions and Importance of Proxy Data for the Mediterranean 77 1.4. Mediterranean Winter Temperature and Precipitation Variability over the Last 500 Years 83 1.5. Connection between the Large-Scale Atmospheric Circulation and Mediterranean Winter Climate over the Last Centuries 98 1.5.1. Major Atmospheric Circulation Patterns Associated with Mediterranean Winter Climate Anomalies 98 1.5.2. Mediterranean Climate Variability since the Mid-Seventeenth Century in Terms of Large-Scale Circulation Dynamics 103 1.5.3. Running Correlation Analysis between the Large-Scale Atmospheric Circulation and Mediterranean Winter Climate over the Last Centuries 105
viii
Contents
1.6. Teleconnection Studies with Other Parts of the Northern Hemisphere 1.7. Mediterranean Winter Temperature and Precipitation Reconstructions in Comparison with the ECHO-G and HadCM3 Coupled Models 1.7.1. Past Climate Variability and its Relations to Volcanism, Solar Activity and GHG Concentrations: Reconstructions and Models 1.8. Conclusions 1.9. Outlook Acknowledgements References
Chapter 2. Relations between Climate Variability in the Mediterranean Region and the Tropics: ENSO, South Asian and African Monsoons, Hurricanes and Saharan Dust 2.1. Introduction 2.2. ENSO Impact on the Mediterranean Climate 2.2.1. ENSO and Eastern Mediterranean (EM) Rainfall 2.2.2. ENSO and the Western Mediterranean Relationship 2.2.3. ENSO and Extreme Mediterranean Rainfall 2.2.4. Transient and Stationary Waves Approach 2.2.5. Possible Coupling Mechanism of ENSO and the Mediterranean 2.3. South Asian Monsoon Variability and the Mediterranean Climate 2.3.1. Mediterranean Climate and the South Asian Rainfall 2.4. African Monsoon Impact on the Climate of the Mediterranean 2.5. Tropical Cyclones’ Impact on the Mediterranean Climate 2.6. Tropical Intrusions into the Mediterranean Basin 2.7. Mediterranean Dust Transport from Sahara 2.8. Conclusions and Outlook 2.8.1. Future Research on ENSO Impact on Mediterranean Climate 2.8.2. Future Research on South Asian Monsoon Variability and the Mediterranean Climate 2.8.3. Future Research on African Monsoon Impacts on the Climate of the Mediterranean 2.8.4. Future Research on Tropical Intrusions into the Mediterranean Basin Acknowledgements References
109
113
117 119 122 125 126
149 149 150 151 154 156 157 157 159 160 161 162 164 165 168 170 170 171 171 172 172
Contents ix
Chapter 3. Relations between Variability in the Mediterranean Region and Mid-Latitude Variability 3.1. Introduction 3.2. Mid-Latitude Modes of Atmospheric Variability and their Impact 3.2.1. The North Atlantic Oscillation Pattern 3.2.2. The Eastern Atlantic and Eastern Atlantic/Western Russia Patterns 3.2.3. The Scandinavian and Blocking Patterns 3.2.4. Other Modes 3.3. Temperature Variability 3.4. Precipitation Variability 3.5. Trends 3.5.1. Temperature Trends 3.5.2. Precipitation Trends 3.5.3. Contributing Factors for Observed Temperature and Precipitation Trends 3.6. Other Important Forcing Factors 3.6.1. Tropical and Extratropical SST 3.6.2. Solar Variability 3.7. Future Outlook Acknowledgements References Chapter 4. Changes in the Oceanography of the Mediterranean Sea and their Link to Climate Variability 4.1. Introduction 4.2. The Forcing of the Mediterranean Sea 4.2.1. Air–Sea Interaction 4.2.2. River Outflow 4.2.3. Exchanges through the Strait of Gibraltar 4.2.4. The Exchange with the Black Sea 4.3. The Mediterranean Overturning Circulation 4.3.1. The Modification of the Atlantic Inflow 4.3.2. Intermediate Water Formation 4.3.3. Deep Water Formation in the Eastern Mediterranean 4.3.4. Deep Water Formation in the Western Mediterranean 4.4. Climatic Changes in the Mediterranean Sea Circulation 4.4.1. Multidecadal Trends in Water Mass Characteristics 4.4.2. Rapid Changes: The Eastern Mediterranean Transient 4.4.3. What Caused the EMT? 4.5. The Impact of Large-Scale Atmospheric Variability on the Mediterranean Sea 4.6. Sea Level Changes in the Mediterranean Sea 4.6.1. Extreme Sea Levels
179 180 182 183 187 189 191 192 193 198 198 202 204 212 212 213 215 218 219
227 227 229 229 231 235 236 239 239 241 243 245 246 246 249 251 260 261 263
x Contents
4.7. Changes in the Wind-wave Field 4.8. Outlook and Future Challenges Acknowledgements References Chapter 5. The Atlantic and Mediterranean Sea as Connected Systems 5.1. Introduction 5.2. Mediterranean Outflow vs. Mediterranean Internal Variability 5.2.1. Description of Numerical Model 5.2.2. Zonal and Meridional Cells 5.2.3. Mean State and Variability 5.2.4. Timescales of Mediterranean Thermohaline Circulation 5.2.5. Decadal Oscillations in the Mediterranean Sea 5.3. The Strait of Gibraltar: A Gate to the Atlantic 5.4. Spreading of Mediterranean Outflow Water in the North Atlantic 5.4.1. Dynamics of Mediterranean Outflow Water 5.4.2. Mediterranean Outflow Water and Stability and Variability of Meridional Overturning Circulation 5.4.3. Modelling of Mediterranean Outflow Water 5.5. Discussion and Future Challenges References Chapter 6. Cyclones in the Mediterranean Region: Climatology and Effects on the Environment 6.1. Introduction 6.1.1. ‘‘Historical’’ Notes 6.2. Mediterranean Cyclones: Data, Methods and Dynamics 6.2.1. Dynamics of Cyclones in the Mediterranean Region 6.2.2. Available Sets of Data 6.2.3. Methodology 6.3. Climatology of Cyclones in the Mediterranean 6.3.1 Characteristics, Sub-Areas of Cyclogenesis, Seasonality and Generation Mechanisms 6.3.2. The Role of Large-Scale Climate Patterns on the Mediterranean Cyclones 6.3.3. Trends 6.4. Weather Patterns and Mediterranean Environment 6.4.1. Precipitation 6.4.2. Strong Winds 6.4.3. Storm Surge 6.4.4 Wind Waves 6.4.5. Landslides
265 268 270 270 283 283 284 285 285 286 287 293 299 303 305 307 310 314 316
325 325 328 329 329 332 333 337 337 342 343 344 345 351 353 357 360
Contents xi
6.5. Conclusions 6.6. Outlook Acknowledgements References
362 364 365 365
Chapter 7. Regional Atmospheric, Marine Processes and Climate Modelling 7.1. Introduction 7.2. Teleconnection Patterns from the Mediterranean Region 7.3. Mediterranean Thermohaline Circulation and its Sensitivity to Atmospheric Forcing 7.4. Sensitivity of the Mediterranean Thermohaline Circulation to Anthropogenic Global Warming 7.5. Current Status of Mediterranean Regional Climate Modelling 7.6. Atmosphere–Sea Coupled Modelling 7.7. Perspectives and Outlooks 7.7.1. High-Resolution Mediterranean Climate Modelling Systems 7.7.2. Development and Validation of Integrated Regional Modelling Systems Acknowledgements References
373 373 376
Chapter 8. The Mediterranean Climate Change under Global Warming 8.1. Introduction 8.2. Global-Coupled Climate Models 8.3. Regional Climate Scenarios 8.3.1. Statistical Downscaling 8.3.2. RCM Simulations 8.3.3. AGCM Simulations with Variable Resolution 8.4. Impacts on Land Surface and Vegetation 8.5. Future Research References
399 399 400 405 405 406 407 412 413 413
Subject Index
417
379 381 386 387 390 391 392 393 393
This page intentionally left blank
Preface
The interest in the Mediterranean Climate is motivated by the convergence of scientific, environmental, social and economic issues in this region, which is a densely populated area under environmental stress and potentially very sensitive to climate change. Presumably, the most critical situation is associated with the availability of water resources, whose further decrease would greatly affect a large fraction of the population living in the region. There are other issues such as extremes which are of much relevance for the area. After the hot summer of 2003, most likely the hottest in the larger Mediterranean area for more than half-amillennium, heat waves are perceived to be a major danger. Floods and landslides, both as a consequence of single intense precipitation events and of periods of persistent precipitation, are a major problem because of the morphology of the territory, with small and steep river basins. The very long coastline, most of it densely populated, could be affected by change of storms and wave regimes. Damage is caused every year by small-scale events such as hail, lightning and tornados. It is important to understand how the large-scale climate variability and change influence the climate of the Mediterranean region in the past, present and the future, considering the effects of the mesoscale features, orography, land–sea interaction and regional mechanisms characterizing it. Finally, the Mediterranean Sea plays a role in the global climate system, as a source of moisture, heat reservoir and being the origin of the salty water exiting at intermediate levels into the Atlantic Ocean. These motivations have already produced several initiatives. The European Geophysical Society and, subsequently, the European Geophysical Union have inserted a Mediterranean Climate session in their annual conference programme since 2003. In 2004, the European Science Foundation supported a workshop on Mediterranean Climate, during which the material included in this book was extensively discussed. A project focused on the Mediterranean Climate, called MedCLIVAR, has been endorsed by the CLIVAR (Climate Variability and predictability) project of the World Climate Research Programme. A specific international project on cyclones that produce high-impact weather in the Mediterranean (MEDEX) was endorsed by the World Meteorological
xiv Preface
Organization, in the frame of the World Weather Research Programme. Besides these, many other initiatives at European and national levels are addressing regional aspects of the Mediterranean Climate and investigating the impacts of its variability and change in future climate scenarios. This book aims to present in a coherent and organic way the results of several of these initiatives and give an updated view of the research recently carried out, with emphasis on the analysis at the Mediterranean basin scale and considering global implications. The book is a multi-authored book, where groups of scientists, co-ordinated by a leading author are responsible for each chapter. It is the result of the cooperation among the scientists involved in the MedCLIVAR project and such cooperation is reflected in the integration and cross-referencing among chapters, where the characteristics of the Mediterranean Climate, its variability and trends are discussed from different perspectives. The focus is on decadal and centennial timescales and on the results available on the impact of future emission scenarios at regional scale. The physical processes responsible for this variability are both local – such as changes in the surface properties and land use – and global – such as changes in the large-scale atmospheric circulation associated with global warming, the North Atlantic Oscillation (NAO), tropical monsoon and El Nin˜o-Southern Oscillation (ENSO). Some limitation on subjects included has been self-imposed. Though paleo data describing longer timescales and interactions involving ecosystems are obviously interesting and related topics, it was preferred to restrict the scope to the physical component in order to have a homogeneous set of chapters. Subjects directly related to seasonal climate prediction were not included as it was felt that they were already addressed by other groups of scientists and weather prediction centres. This book consists of an introduction and eight chapters. The introduction describes the main characteristics of the Mediterranean Climate and presents briefly the main factors characterizing its dynamics and its role on the global climate. The first chapter (Luterbacher et al.) reviews the past Mediterranean Climate variability covering the last few centuries. It describes the regional coverage, the potential and limitations of the available instrumental data, documentary evidence and natural proxies. It, moreover, presents yet unexplored archives (marine and land) and their potential for past climate reconstructions. The Mediterranean Climate, because of its location, is under the influence of both mid-latitude and tropical dynamics (e.g. NAO on one side and ENSO and Monsoons on the other side). These links are discussed separately in the second and third chapter (Alpert et al. and Trigo et al.), respectively. The fourth chapter (by Tsimplis et al.) describes the variability of the Mediterranean Sea circulation and sea level (in relation with that at the global scale) and their relations to largescale climate pattern. The material includes also the description of abrupt changes of the Mediterranean Sea circulation and their mechanisms. The fifth
Preface
xv
chapter (Artale et al.) analyses the variability of the Mediterranean outflow into the Atlantic across the Gibraltar Strait and its role on the Atlantic Meridional Overturning Circulation. The sixth chapter (Lionello et al.) describes the climatology of cyclones in the Mediterranean region and of the extreme events associated with them, such as floods, heavy rains, wind waves and storm surges. Regional climatic processes, air–sea interaction and their possible effect on remote areas (e.g. the link between Mediterranean Sea Surface Temperature and the precipitation in the Sahel region) are discussed in the seventh chapter (Li et al.). The final eighth chapter (Ulbrich et al.) is devoted to the analysis of the effect of future emission scenario on the Mediterranean Climate considering the results of multi-model and ensemble simulations. This book aims to review the research on the Mediterranean Climate and, at the same time, to provide both an introduction and a reference for researchers. It covers topics typical of Climatology, Climate history, Meteorology, Oceanography, Environmental Science but the information provided here would also be useful for research in agriculture, social and economic studies. It is addressed to scientists and students interested in the Mediterranean climate and environment. Some topics have interesting connections to nearby regions: Northern Atlantic, West Africa, Central and Eastern Europe. Each chapter contains a summary meant to provide information to policy makers, researchers from other fields, and in general, to a wide audience without a technical expertise on climate. P. Lionello P. Malanotte-Rizzoli R. Boscolo
This page intentionally left blank
Introduction
The Mediterranean Climate: An Overview of the Main Characteristics and Issues P. Lionello,1 P. Malanotte-Rizzoli,2 R. Boscolo,3 P. Alpert,4 V. Artale,5 L. Li,6 J. Luterbacher,7 W. May,10 R. Trigo,8 M. Tsimplis,9 U. Ulbrich11 and E. Xoplaki7 1
Department of Material Sciences, University of Lecce, Italy (
[email protected]) 2 Massachusetts Institute of Technology, USA (
[email protected]) 3 ICPO, UK and Spain (
[email protected]) 4 Tel Aviv University, Israel (
[email protected]) 5 ENEA, Roma, Italy (
[email protected]) 6 Laboratory of Dynamical Meteorology CNRS, Paris, France (
[email protected]) 7 Institute of Geography and NCCR Climate, University of Bern, Switzerland (
[email protected],
[email protected]) 8 University of Lisbon, Portugal (
[email protected]) 9 National Oceanography Centre, Southampton, UK (
[email protected]) 10 Danish Meteorological Institute, Copenhagen, Denmark (
[email protected]) 11 Freie Universita¨t Berlin, Germany (
[email protected])
1. The Mediterranean Region: Climate and Characteristics The Mediterranean Region has many morphologic, geographical, historical and societal characteristics, which make its climate scientifically interesting. The purpose of this introduction is to summarize them and to introduce the material extensively discussed in the succeeding chapters of this book. The connotation of ‘‘Mediterranean climate’’ is included in the qualitative classification of the different types of climate on Earth (e.g. Ko¨ppen, 1936) and it has been used to define the climate of other (generally smaller) regions besides that of the Mediterranean region itself. The concept of ‘‘Mediterranean’’ climate is characterized by mild wet winters and warm to hot, dry summers and may occur on the west side of continents between about 30 and 40 latitude. However, the presence of a relatively large mass of water is unique to the actual Mediterranean region. The Mediterranean Sea is a marginal and semi-enclosed
2 Mediterranean Climate Variability
sea; it is located on the western side of a large continental area and is surrounded by Europe to the North, Africa to the South and Asia to the East. Its area, excluding the Black Sea, is about 2.5 million km2; its extent is about 3,700 km in longitude, 1,600 km in latitude. The average depth is 1,500 m with a maximum value of 5,150 m in the Ionian Sea. It is surrounded by 21 African, Asian and European countries. The Mediterranean Sea is an almost completely closed basin, being connected to the Atlantic Ocean through the narrow Gibraltar Strait (14.5 km wide and less than 300 m deep). These morphologic characteristics are rather peculiar. In fact, most of the other marginal basins have much smaller extent and depth or they are connected through much wider openings to the ocean. An example of the first type is the Baltic. Examples of the second type are the Gulf of Mexico and the Arabian Sea. The closest analogue to the Mediterranean is possibly the Japan Sea, which, however, does not have a similar complex morphology of basins and sub-basins and is located on the eastern side of the continental area. A specific characteristic of the Mediterranean region is its complicated morphology, due to the presence of many sharp orographic features, the presence of distinct basins and gulfs, islands and peninsulas of various sizes (Fig. 1). High mountain ridges surround the Mediterranean Sea on almost every side and tend to produce much sharper climatic features than expected without their existence. The highest ridge is the Alps, reaching a maximum high of 4,800 m, which contains permanent glaciers and presents a thick and extended snow cover in winter. Islands, peninsulas and many regional seas and basins
Figure 1: Orography and Sea-depth of the Mediterranean region.
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
3
determine a complicated land–sea distribution pattern. These characteristics have important consequences on both sea and atmospheric circulation, because they determine a large spatial variability and the presence of many subregional and mesoscale features. The oceanic topography is similarly complicated with deep basins linked through much shallower straits. The Mediterranean Sea circulation is characterized by sub-basin scale gyres, defined by the geometry and topography of the basin, and dense water formation processes, which are responsible for its deep circulation (Tsimplis et al., Chapter 4 of this book).
Geographical Elements in the Map Straits (denoted with white arrows) 1-Strait of Gibraltar 2-Strait of Sicily 3-Strait of Otranto 4-Cretan Strait (West) 5-Cretan Straits (East) 6-Dardanelles 7-Bosporus Strait Mountains -Alps -Anatolian mountains -Apennines -Atlas mountains -Balkans -Dinaric Alps -Pyrennees Lakes -Sea of Galilee -Dead Sea
Gulfs (denoted with circles) 1-Gulf of Lion 2-Gulf of Genoa 3-Gulf of Venice 4-Gulf of Sirte Islands -Balearic Islands -Corsica -Crete -Cyprus -Rhodes -Sardinia -Sicily Peninsulas -Balkan peninsula -Crimea -Iberian peninsula -Italian peninsula
Seas and Basins (denoted with boxes) 1-Alboran Sea 2-Algerian basin 3-Tyrrhenian Sea 4-Adriatic Sea 5-Ionian Sea 6-North Aegean Sea 7-Cretan Sea 8-Cyclades Plateau 9-Levantine basin 10-Black Sea 11-Red Sea Rivers (mouths are denoted with black arrows) -Ebro -Nile -Po -Danube -Jordan Others -The Negev desert
Figure 2: Map with labels denoting most relevant geographical features of the Mediterranean region.
4 Mediterranean Climate Variability
The atmospheric circulation is strongly affected by the complex land topography which plays a crucial role in steering air flow, so that energetic mesoscale features are present (Lionello et al., Chapter 6 of this book). The large environmental meridional gradient is shown by the transition from hot and arid regions to humid mountain climate and permanent glaciers in about 2,000 km. Furthermore, strong albedo differences exist in south–north directions (Bolle, 2003). Figure 2, which provides a reference for the geographic features mentioned in this book, shows the large amount of details involved in the description of the mesoscale forcings in this region. Because of its latitude, the Mediterranean Sea is located in a transitional zone, where mid-latitude and tropical variability are both important and compete (Alpert et al., and Trigo et al., Chapters 2 and 3 of this book, respectively). Thus, from a Koppen classification perspective, the northern part of the Mediterranean region presents a Maritime West Coastal Climate, while the Southern part is characterised by a Subtropical Desert Climate. Further, the Mediterranean climate is exposed to the South Asian Monsoon in summer and the Siberian highpressure system in winter. The southern part of the region is mostly under the influence of the descending branch of the Hadley cell, while the Northern part is more linked to the mid-latitude variability, characterized by the NAO (North Atlantic Oscillation) and other mid-latitude teleconnection patterns (e.g. Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2003, 2004; Hoerling et al., 2004; Hurrell et al., 2004). An important consequence is that the analysis of the Mediterranean climate could be used to identify changes in the intensity and extension of global-scale climate patterns, such as NAO, ENSO (El Nin˜o Southern Oscillation) and the Monsoons. The teleconnections in the Mediterranean region present a large amount of both spatial variability (ranging from synoptic to mesoscale) and time variability (with a strong seasonal cycle modulated on multi-decadal to centennial time scales, as described in the Chapters from 1 to 6 of this book). Moreover it is important to consider the role of the Mediterranean Sea as heat reservoir and source of moisture for surrounding land areas; as source of energy and latent heat for cyclone development (Lionello et al., Chapter 6), and its possible effect on remote areas (such as the Sahel region in, Li et al., Chapter 7 of this book) and on the Atlantic overturning circulation (Artale et al., Chapter 5). Another important characteristic of the Mediterranean region is the large amount of climate information from past centuries (Luterbacher et al., 2004, Chapter 1 of this book; Guiot et al., 2005; Xoplaki et al., 2005). This characteristic is shared with other European regions, but apart from them, is presently unique on the global scale and has not yet been fully exploited. The continuous presence of well-organized local states and the long tradition of scholarship and natural science produced documentary proxy evidence, which allows the
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
5
reconstruction of some aspects of climate since the Roman period and possibly further back in time. Some millennial-long climate series have already been reconstructed (e.g. of the freezing of the Venetian lagoon and of storm surge in Venice; Camuffo, 1987, 1993). This availability of documentary evidences is complemented with natural proxies (tree ring data, corals, etc., Felis et al., 2000; Touchan et al., 2003, 2005) as well as with remarkably long observational records (associated with old universities and observatories of municipalities, kingdoms and counties) mostly on the central and western-European part of the Mediterranean region (e.g. Buffoni et al., 1999; Barriendos et al., 2002; Camuffo, 2002; Maugeri et al., 2002; Rodrigo, 2002). On the basis of documentary and/or natural proxies it has been possible to obtain multi-centennial regional temperature and precipitation reconstructions (e.g. Guiot et al., 2005; Luterbacher et al., 2004; Mann, 2002; Till and Guiot, 1990; Touchan et al., 2003, 2005) allowing to study of past climate variability, trends, uncertainties and to compare the Mediterranean region with other areas. This rich data gives a unique opportunity for reconstruction of climate (including extremes) in past historical and recent instrumentally developed times. An important characteristic of the Mediterranean region is the emergence of highly populated and technologically advanced societies since, at least, 2000 BC. Because of the demographic pressure and exploitation of land for agriculture, the region presents, since ancient times, important patterns of land use change and important anthropic effects on the environment, which are themselves interesting research topics. For example, it has been suggested that the albedo change due to the change in vegetation since Roman times significantly alters the atmospheric circulation over northern Africa and the Mediterranean Sea, so that deforestation around the Mediterranean during the last 2,000 years may be a major factor in the dryness of the current climate in these regions (Reale and Dirmeyer, 2000; Reale and Shukla, 2000). The importance of deforestation in the Mediterranean region has been confirmed by other modelling studies suggesting that lower plant evapotranspiration and lower evaporation from soils, due to erosion, are likely to reduce precipitation in summer (Du¨menil-Gates and Liess, 2001). The Mediterranean Sea general circulation has been described through a series of observational programmes and modelling studies over the past 20 years (e.g. POEM, PRIMO, WMCE, EU/MAST/MTP I and EU/MAST/MTP II). The modern reconstruction of the basin-wide general circulation (POEM Group, 1992; Millot, 1999) and its variability results much more complicated than that described before (Ovchinnikov, 1966), spanning over multiple scales in space (from the basin-scale to the sub-basin and mesoscale) and in time (from the seasonal to the interannual and decadal variability). Fundamental components of the basin-scale circulation are three major thermohaline cells. The first
6 Mediterranean Climate Variability
one is the ‘‘open’’ circulation cell that connects the eastern to the western Mediterranean and is associated with the inflow of Atlantic Water at the Gibraltar in the surface layer and the outflowing return flow of Levantine Intermediate Water (LIW) in the intermediate layer below. The others are two meridional vertical cells confined to the eastern and western Mediterranean basins. They are driven by localized deep convective events, which occur in the Northern Mediterranean areas, leading to the formation of dense water masses, which spread in the deepest layers, with subsequent upwelling and return flow at the intermediate layer into the convection region. The importance of localized convection processes is determined by air–sea interaction and long-term preconditioning. Intense cooling and evaporation over restricted areas in the north-western Gulf of Lion, the southern Adriatic Sea and, in the 1990s, the Aegean/Cretan Sea control the formation of dense waters filling the bottom of the basin. The western and eastern sub-basins are disconnected at deep levels, hence their thermohaline circulations are independently driven by the respective sources. The eastern Mediterranean thermohaline circulation is a closed cell endowed with multiple equilibria. Analogous observational evidence and related modelling studies, for the Western Mediterranean are lacking. Intense evaporation in the Levantine basin determines the formation of LIW which is part of the open thermohaline cell constituted by two branches: Atlantic Water entering at Gibraltar and making its way to the Levantine, being transformed into LIW by intermediate convection processes (mainly in the Rhodes gyre area), and returning all the way to Gibraltar, where it finally exits forming the North Atlantic salty water tongue.
2. Regional Processes and Links to the Global Climate The climate of the Mediterranean region is to a large extent forced by planetary scale patterns. The time and space behaviour of the regional features associated with such large-scale forcing is complex. Orography and land–sea distribution play an important role establishing the climate at basin scale and its teleconnections with global patterns. In fact, the complexity of the basin topographic structures, which have been described in the previous section, implies the presence of mesoscale features and inter-seasonal variability in patterns that would be otherwise much more homogeneous and persistent. Most studies consider winter and summer regimes, while characterization of spring and autumn is more uncertain, revealing, presumably, the transient nature of these two seasons in the Mediterranean region. The large-scale mid-latitude atmospheric circulation exerts a strong influence on the cold season precipitation over the Mediterranean, though the strength
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
7
of the relation varies across the region and depends on the considered period (see Chapters 1, 2 and 3). The largest amount of studies refer to the role of the NAO, which determines a large and robust signal on winter precipitation, which is anti-correlated with NAO over most of the western Mediterranean region (Hurrell, 1995; Dai et al., 1997; Rodo et al., 1997; Xoplaki, 2002; Trigo et al., 2004). This strong link is due to the control exerted by NAO on the branch of the storm track affecting the Mediterranean, mainly in its western part. Besides the NAO, other patterns influence the Mediterranean climate (Corte-Real et al., 1995; Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2004). The role of the Mediterranean Sea itself as source of moisture and the subsequent eastward advection by the atmospheric circulation imply a more complex picture for the Eastern Mediterranean, where the EA (East Atlantic) pattern plays an important role (Krichak et al., 2002; Fernandez et al., 2003). In general, EA describes much of the precipitation anomalies in the whole basin that cannot be ascribed to the NAO (Quadrelli et al., 2001). Moreover, in the central Mediterranean the Scandinavian pattern has a strong influence (e.g. Xoplaki, 2002). The influence of ENSO in the North Atlantic–European area has been identified mostly in winter during its extreme events (Pozo-Va´zquez et al., 2001). In fact, ENSO has been found to play an important role in winter rainfall in the eastern Mediterranean, where the role of NAO is weak (e.g. Yakir et al., 1996 and Price et al., 1998). Specifically, higher/lower than normal precipitation in Israel have been shown for El Nin˜o/La Nin˜a years. This is associated with the meridional shift of the jet above the Eastern Mediterranean region, observed during El Nin˜o/La Nin˜a years. In the western Mediterranean, as, in general, for the North Atlantic/European regions, it is difficult to identify the ENSO signals by using common statistical techniques mainly due to their spiky nature with respect to the dominating mid-latitude dynamics (Rodo´, 2001). Their importance for the Mediterranean climate is not clear, though it might be large for selected intervals and vanish elsewhere (Rodo´ et al., 1997; Rodo´, 2001; Mariotti et al., 2002a). The most robust link is that of western Mediterranean-autumn-averaged rainfall, which has a significant positive correlation with ENSO. A weaker correlation, with opposite sign, is present in spring, but it is confined to a small region in Spain and Morocco (Mariotti et al., 2002a). In summer, when the advection of moisture from the Atlantic is weaker and the Hadley cell moves northward and its strength diminishes, there are evidences of connections (stronger in the eastern Mediterranean and at the North African coast) with the Asian and the African monsoons. Rodwell and Hoskins (1996) pointed at the linkage between the appearance of the semi-permanent subsidence region over the eastern Mediterranean and the onset of the monsoon. A consequence could be that the very dry summertime climate of the Eastern Mediterranean and the surrounding lands may be strongly related to the
8 Mediterranean Climate Variability
characteristics of the Asian monsoon regime. Ziv et al. (2004), in their study of the summer atmospheric circulation, pointed out also to the role of the Hadley cell over eastern North Africa, connecting the eastern Mediterranean with the African Monsoon. Such influence does not extend to the Northern Mediterranean rainfall variability, which in summer has been shown to be related with the EA Jet pattern (Du¨nkeloh and Jacobeit, 2003). The influence of NAO on the Mediterranean winter temperature is smaller than that on precipitation. Above the North-western Mediterranean region the spatial distribution of correlation with temperature has a positive feature which is weaker than the negative one of the correlation with precipitation. The effect of NAO has been found to be non-linear and non-stationary (e.g. Pozo-Va´zquez et al., 2001). Moreover, the strong control exerted by NAO on cloud cover and surface radiation balance over the Mediterranean implies the appearance of asymmetric NAO impact patterns for maximum and minimum temperature (Trigo et al., 2002). Other studies (Sa´enz et al., 2001; Frı´ as et al., 2005) suggest that the variability of monthly mean winter temperature over the western part of the Mediterranean basin is mainly controlled by the variability of the EA pattern, with the NAO only playing a secondary role. The reason behind this is that this variability is controlled by sensible heat transport by mean stationary waves, with eddy heat fluxes playing a less significant role. However, the influence of EA cannot be extended to the eastern part of the basin (Hasanean, 2004). Mediterranean summer temperatures are not related with NAO neither with monthly indices of other large-scale patterns. Rather, generally, warm Mediterranean summers are connected with a main mode, characterized by strong positive geopotential anomaly covering large parts of Europe including the Mediterranean area, associated with blocking conditions, subsidence, stability, a warm lower troposphere, small pressure gradients at sea level as well as above-normal Mediterranean SST (Xoplaki et al., 2003). An important role of the Mediterranean Sea on the climate of other regions is associated with the connection between the Mediterranean SST (Sea Surface Temperature) and the Sahel precipitation. The dry and hot summers in the western Mediterranean and the Monsoon regime over West Africa, are correlated through a positive mechanism. When the Mediterranean SST is higher than normal, moisture from the western Mediterranean is fed into the Sahel favouring precipitation (Semazzi and Sun, 1997; Rowell, 2003). In turn, abundant rainfall in the Sahel increases the high pressure upstream the western Mediterranean via Rossby waves (Rodwell and Hoskins, 1996), enhancing the subsidence in the western Mediterranean, favouring the penetration of moisture air from the Atlantic into the Sahel and increasing further the rainfall in this region. The effect of SST anomalies in summer has been investigated with a relatively
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
9
high-resolution (T159L40) version of the ECMWF model in relation with the situation which took place in summer 2003 (Jung et al., 2005) confirming that anomalously warm Mediterranean SSTs are associated with an intensification of Sahel rainfall. Also other modelling studies suggest that changes of SST in the Mediterranean can have consequences for atmospheric circulation, affecting relatively remote regions (Li, 2005). The dynamics of the Mediterranean climate includes also the identification of processes acting internally to the region. With respect to the moisture balance, the western Mediterranean Sea represents a source for the surrounding land areas and for the eastern part of the basin, as the moisture released by evaporation is redistributed by the atmospheric circulation (Fernandez et al., 2003). These processes could play a role in coupled atmosphere–ocean variability modes, which could be important for the regional hydrological cycle, and the water budget of the eastern Mediterranean region. In fact, wet and dry winter months in the Eastern Mediterranean region are characterized by circulation patterns with north-westerly and north-easterly air flow in the lower troposphere above the Eastern Mediterranean (Krichak and Alpert, 2005a). Regional weather regimes are a basic element of the Mediterranean climate. These regimes are characterized by energetic mesoscale features, several cyclogenesis areas (e.g. Alpert et al., 1990; Trigo et al., 1999; Lionello et al., 2002) and are generally characterized by shorter life-cycles and smaller spatial scales than extra-tropical cyclones developed in the Atlantic. Extremely diversified classes of cyclones are present in the Mediterranean region, since it presents geographic factors that can substantially influence the cyclogenesis mechanisms. A tentative list, based partially on the mechanisms producing cyclogenesis and partially on the geographical characteristics, would include lee cyclones, thermal lows, small-scale hurricane-like cyclones, Atlantic systems, African cyclones, Middle East lows. Though NAO plays a basic role and many Mediterranean cyclones are triggered by synoptic systems passing over Central and Northern Europe along the Northern Hemisphere storm track, also other teleconnection patterns with centres of action located closer to or above Europe have to be considered (e.g. Rogers, 1990, 1997). Studies of the variability of the Mediterranean Sea circulation have aimed to identify the forcings that are responsible for the variability of the circulation patterns, especially in relation to long-term trends in water mass characteristics, sea level changes and the EMT (Eastern Mediterranean Transient) event during which the source of the Eastern Mediterranean deep water temporarily moved from the Southern Adriatic to the Aegean/Cretan Sea. Various factors have been identified by different authors, without reaching definitive conclusions on the dominant mechanism for the EMT. The wind stress climatology over the eastern Mediterranean was quite different in the 1980s and the 1990s
10 Mediterranean Climate Variability
(Samuel et al., 1999) and this forcing variability may have induced the observed important change in the eastern Mediterranean upper thermocline circulation which affected the LIW pathways (Malanotte-Rizzoli et al., 1999). Internal mechanisms may equally be at play, such as an internal redistribution of salt in the eastern basin (Roether et al., 1996). The construction of the Nile dam and the diversion of Russian rivers are argued to have had an impact in the EMT event (Boscolo and Bryden, 2001; Skliris and Lascaratos, 2004) as well as to have increased the overall salinity of the whole Mediterranean basin (Rohling and Bryden, 1992). As regard air–sea fluxes, severe heat loss in two winters in the early 1990s affected the surface heat budget and increased surface buoyancy loss (Josey, 2003). The Black Sea outflow plays a role in the Aegean hydrographical conditions. The induced freshening of the surface layer of the North Aegean hinders dense water production in that region (Plakhin, 1972); where, however, particular events of massive dense water production have been recorded in 1987 and 1993 (Theocharis and Georgopoulos, 1993; Zervakis et al., 2000). Evidence suggests that when dense water production takes place in the North Aegean, it is partly induced by a decrease of the Black Sea buoyancy input (Zervakis et al., 2000; Nittis et al., 2003). Whether the dense water production of the North Aegean affects the Cretan Deep Water production (and thus the EMT) is still a matter of investigations. The links between these mechanisms and large-scale patterns of atmospheric climate variability still needs to be properly assessed. A possible feedback of the Mediterranean Sea on the global climate could derive from an effect of air–sea interaction in the Mediterranean region on the Atlantic Meridional Overturning Circulation (MOC) via the Mediterranean outflow across the Gibraltar Strait, which determines the presence of a wellknown tongue of very salty water in the entire Northern Atlantic at intermediate depths (1,000–2,500 m). This water introduces an important signature in the salinity field and has potentially important large-scale climate implications (Chapter 5 of this book). Two mechanisms are presently envisioned through which the salty Mediterranean water may influence the convective cells in the Greenland and Labrador seas where North Atlantic Deep Water (NADW) is formed. The first mechanism involves a direct advective pathway from the Strait of Gibraltar to the polar seas (Reid, 1994). The second mechanism involves a progressive lateral mixing of the Mediterranean water with Atlantic intermediate water and entrainment in the North Atlantic current that reaches the polar seas (Lozier et al., 1995; Potter and Lozier, 2004). Both mechanisms imply a significant role for the Mediterranean water in preconditioning the surface water column of the convective cells, thus increasing the volumes of newly formed NADW, and increasing the stability of the MOC. The interaction of the Mediterranean outflow with the thermohaline circulation of the North Atlantic
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
11
raises the possibility for feedback mechanisms, eventually active both at decadal and millennial time scales, involving the North Atlantic, the Mediterranean and the overlying atmosphere, which have potentially important climatic implications (Rahmstorf, 1998; Artale et al., 2002).
3. Climate Trends and Change at Regional Scale Precipitation and temperature in the Mediterranean during the 20th century show significant trends. Negative precipitation trends have been present at different time and space scales (e.g. Folland et al., 2001; New et al., 2001). Giorgi (2002a) found negative winter precipitation trends over the larger Mediterranean land area for the 20th century. However, sub-regional variability is high and trends in many regions are not statistically significant in view of the large variability (e.g. Xoplaki, 2002). Giorgi (2002a) analysed also the surface air temperature variability and trends over the larger Mediterranean land area for the 20th century based on gridded data of New et al. (2000). He found a significant warming trend of 0.75 C in one-hundred years, mostly from contributions in the early and late decades of the century. Slightly higher values were observed for winter and summer. The structure of climate series can differ considerably across regions showing variability at a range of scales. Over most of western Mediterranean for instance, warming has been mainly registered in two phases: from the mid-1920s to 1950s and from the mid-1970s onwards (e.g. Brunet et al., 2001; Galan et al., 2001; Xoplaki et al., 2003). Moreover, the availability of documentary and natural proxies in the Mediterranean region has allowed constructing seasonally resolved temperature and precipitation maps over the Mediterranean area for more than 500 years with associated uncertainties (Luterbacher et al., 2004; Pauling et al., 2005; Xoplaki et al., 2005). In this context, the analysis of winter temperature and precipitation reveals that the recent winter decades (end of twentieth, beginning of the twenty-first century) were the warmest and driest, in agreement with recent findings from Europe and the Northern Hemisphere. It is obviously important to investigate the future evolution of these trends and produce reliable model simulations. Important environmental changes have been observed in the Mediterranean Sea circulation during the last decades. Warming trends have been observed both in deep and intermediate water (e.g. Bethoux et al., 1990, 1998). Sea level has increased in line with the mean estimated global value (1.8 mm/year) till the 1960s, but it has subsequently dropped by 2–3 cm till the beginning of the 1990s (Tsimplis and Baker, 2000). During the last decade of the 20th century, sea level has increased 10 times faster than on global scale. The EMT, already discussed in the previous section, is a major change which has characterized the
12 Mediterranean Climate Variability
eastern Mediterranean deep water formation with a transition in the structure of its closed internal thermohaline cell. Historically, the eastern Mediterranean thermohaline circulation has been driven by a deep convection site for dense water formation localized in the southern Adriatic Sea (Roether and Schlitzer, 1991). This was, in fact, the situation in the 1980s. Between 1987 and 1991, however, the ‘‘driving engine’’ of the thermohaline circulation shifted to the Southern Aegean/Cretan Sea. In 1995, almost the 20% of the deep and bottom waters (below the 1,200 m) of the eastern Mediterranean were found to be replaced by the much more dense waters that spread out from the Aegean Sea through the Cretan Arc Straits (Roether et al., 1996; Malanotte-Rizzoli et al., 1999). This major climatic event, better known as the Eastern Mediterranean Transient (EMT), ceased after 1997. In fact, in 1998 the dense waters of Aegean origin were no longer dense enough to reach the bottom layer and the Adriatic Sea regained its role as primary source of dense water (Theocharis et al., 2002; Manca et al., 2003). This observational evidence has led to postulate different equilibria for the thermohaline circulation, which have also been found through box models of the basin (Ashkenazy and Stone, 2003). The implications of these variations for the Mediterranean environment have not been resolved yet. Quantifying and understanding climatic changes at the regional scale is one of the most important and uncertain issues within the global change debate. To date, projections of regional climate changes for the 21st century have been based on coupled atmosphere–ocean global climate model (AOGCM) simulations of the climate system response to changes in anthropogenic forcing (e.g. Kattenberg et al., 1996; Cubasch et al., 2001). An important step towards the understanding of regional climatic changes and impacts is the assessment of the characteristics of natural climate variability and of the AOGCM performance in reproducing it (Giorgi, 2002a,b). Probably, the Mediterranean is one of the few regions where most global simulations carried out with different models give relatively consistent climate change signals. In fact, the various model simulations of the anthropogenic effect on climate tend to agree in their prediction in the Mediterranean region a temperature increase larger than the global average and a large precipitation decrease in summer, but controversial in winter, because of differences among models and between western and eastern areas (see Chapter 8 of this book). In general, in the Mediterranean region, climate change simulations produce a signal in the range from þ 3 to þ 7 K for temperature and from 40 to þ20% for precipitation (Giorgi and Francisco, 2000a,b) in one century. Values obviously depend on the emission scenario. Simulations with five different GCM predict for a A2 scenario an increase in the range 4–5 K in winter and 6–7 K in summer, which in a B2 scenario are reduced to 1 and 2 K, respectively (Parry, 2000). The climate change signal in winter precipitation depends critically on the
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
13
northward deviation of the storm track associated with the shift and intensification of the NAO predicted by some simulations (Ulbrich and Christoph, 1999), which produce decreased and increased precipitation in the southeastern and northwestern Mediterranean region, respectively (Parry, 2000). In summer, all models tend to agree predicting a drier climate with a reduction of precipitation as large as 50% (Parry, 2000). The spatial resolution used for most global climate simulations does not describe adequately the basins that compose the Mediterranean Sea and the mountain ridges surrounding it. The characteristic structures of the Mediterranean region can be identified on the model land–sea mask and surface elevation grid, only if the grid cell size is at least smaller than 50 km. Additionally, even a finer grid is required for describing surface winds and precipitation, whose spatial variability involves scales smaller than 10 km. The information of most environmentally relevant quantities (like temperature, winds, precipitation) in the Mediterranean region is characterized by large spatial variability, as well as high seasonality. Therefore, to extract this from current Global Circulation Models requires some type of ‘‘regionalization’’ meaning ‘‘downscaling’’ with nested Regional Climate Model (RCM), (e.g. Giorgi and Mearns, 1991; Giorgi et al., 1992, 2001; Marinucci and Giorgi, 1992; Machenhauer et al., 1996, 1998; Gonza´lez-Rouco et al., 2000) or the use of global model with variable grid resolution (e.g. the ARPEGE model, De´que´ and Piedelievre, 1995; Gibelin and De´que´, 2003). Many studies, research programmes and European projects have analysed the problem of climate simulation over Europe, but the Mediterranean Sea was most of the times included only partially or in the southern part of the domain, where the model response was likely to be strongly affected by the boundary conditions. Moreover, the quality and reliability of these simulations is currently the object of discussion. It is well known that such RCMs simulations are affected by errors due both to their own dynamics and to those of the global model that provide the boundary conditions. (Christensen et al., 1997; Machenhauer et al., 1996, 1998) and only increasing the resolution itself does not ensure more realistic regional climate scenarios. Regionalization studies confirm the warming in the Mediterranean region predicted by AORCM, stronger is summer than in winter, over land than over sea, and increasing with the intensity of the radiative forcing produced by the emission scenario (De´que´ et al., 1998; Giorgi et al., 2004). The strong reduction of summer precipitation is confirmed using RCM, but the uncertainty on winter values remains high, as the extension of the north western Mediterranean area with higher precipitation depends on the global simulation used to provide the boundary conditions (Raisa¨nen et al., 2004). Recent model simulations (Chapter 7) in which a Mediterranean Sea General Circulation Model has been driven by a climate change scenario for the end of
14 Mediterranean Climate Variability
the 21st century show large possible changes of the Mediterranean Sea circulation. (Somot et al., 2005), with a 3 K temperature and 0.43 psu salinity increase at the sea surface. The lower part (from 500 m depth to the bottom) of the Mediterranean Sea presents a lower increase (0.9 K for temperature 0.18 psu for salinity). The stronger stratification corresponds to a weaker and shallower Mediterranean Thermohaline Circulation. Trends of weather extremes and their changes in future climate scenarios are controversial. During the second half of the 20th century, there is a welldocumented trend showing the reduction of overall winter precipitation in the Mediterranean basin related with significant decrease of intense cyclones (Trigo et al., 2000). At the same time an increase of the relative frequency of torrential rains has been suggested but only for some areas of the western Mediterranean region (Alpert et al., 2002). There are consistent indications of a negative present trend in cyclones, which simulations suggest to continue in future climate scenarios (Lionello et al., 2002), but no clear change of frequency of extreme cyclones has been identified. Critical coastal areas could be heavily affected by changes of marine storminess, extreme storm surge and wind waves events (e.g. the northern Adriatic Sea; Lionello et al., 2003; Lionello, 2005) with large impacts on societies in the Mediterranean region. Because of the difficulty to resolve regional-scale features in global climate simulations, the identification of teleconnections with large-scale circulation patterns is a basic tool for the prediction of future climate extremes (Mun˜oz-Dı´ az and Rodrigo, 2004). Moreover, the regional factors (e.g. Mediterranean Sea surface temperature, moisture content of the air column) are expected to play an important role.
4. Socio-environmental Aspects At present, about 400 million people live in the countries around the Mediterranean Sea and about 145 million (that is 34% of total) people live in the coastal region. The largest number of people live in Italy (32 million), Egypt (24 million), Turkey (11 million), Algeria (10 million) and Greece (9 million). The largest fraction of the total coastline (46,000 km) is in Greece (15,000 km), followed by Italy (8,000 km), Croatia (6,000 km) and Turkey (5,000 km). This densely populated area has large economic, cultural and demographic contrasts1. There are approximately 10-fold differences in GDP (Gross Domestic Product) between the largest economies of the European Union countries and Middle East nations, and a 3- to 6-fold difference in the GDP per-capita between Western European countries and the other nations. The largest economies are those of 1 Information on economic situation, water resources, demographic trends and environmental risks are available at the web page of ‘‘Le plan Bleu’’: http://www.planbleu.org
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
15
France and Italy (about 1,500 and 1,200 trillion $ GNP, respectively). At the same time 8 countries have less than 20 trillion $ GNP. While the same two countries have a per-capita GNP larger than 20,000 $, there are 6 countries with less than 2,000 $2. Demographic trends are also quite different in the area. European countries (also including non-EU nations) are close to null growth and expected to stabilize or even decrease their population, while North African and Middle East countries are growing and are expected to double their population by mid 21st century. In contrast with European Countries, urbanization in most African and Middle East Nations is an ongoing process that is changing the socio-economic structures of these regions. In the second half of the 20th century, a 5- to 10-fold increase in population of large towns has been common for African and Asian countries, with respect to the lower than 2-fold increase of southern European countries. At present, Cairo (which increased 4-fold since 1950) and Istanbul (which increased 8-fold) are the twentieth and twenty-second largest city of the globe, respectively. Population increase is exerting a demographic pressure with an estimated increase from 420 million in 2000 to 530 million in the year 2025 (Tsiourtis, 2001). Migration, caused by soil degradation or even desertification is a very important issue in the Mediterranean area, the southern part of which is mostly dry and vulnerable to climate change. Desertification is not only caused by climate, but also by decisions at all levels in society regarding land use. In fact, natural and human systems are vulnerable to the scarcity and irregular availability of water resources that characterize the region’s climate. Presently, 60% of water-poor world population live in the Mediterranean region (i.e. disposing of less than 1,000 m3 per-capita per year) concentrated in the eastern and southern countries. In fact, about 1,200 km3 are available per year in the whole Mediterranean region (about 50% of them are truly exploitable) and they are very unevenly distributed, so that 71, 20 and 9% of water resources are present in the northern, eastern and southern countries, respectively. Presently 30 million people do not have access to safe water, and structural water shortages are expected to affect 60 million people in 2025. This situation might become very critical if climate change would imply a reduction in precipitation. At present, agriculture still constitutes a major economic activity in the region, corresponding to 20% of the GNP of southern Mediterranean countries, precisely those that are most affected by the variation in water availability. Roughly 50% of the total land area is used for agriculture, absorbing over 80 and 60% of total water demand in the African and European countries surrounding the Mediterranean Sea, respectively. Critical situations of water shortages and extended droughts are mostly due to high values of seasonal and year-to-year 2
All data refer to year 1998
16 Mediterranean Climate Variability
variability in precipitation. In the Mediterranean region, where most of the annual rainfall occurs during the winter half-year, severe water deficits can occur during the growing season even when there is sufficient annual precipitation, so that the impact of temperature and precipitation on most crop yields is mostly dependent on changes in the seasonal cycle of these parameters, rather than on fluctuation in their annual average value (Ro¨tter and van de Geijn, 1999). This is especially true in the eastern Mediterranean and Near East (e.g. Tarawneh and Kadioglu, 2003). The hot and dry spell of summer 2003 produced noticeable damage to agriculture in Spain, Italy and France, where reduction was largest for fodder ( 30, 40, and 60% respectively) and maize ( 10, 25, 30%, respectively), but also heavily affected wheat and potatoes. The generalized dry and hot conditions throughout Europe also had a major impact in increased mortality in the livestock and poultry stocks. The combined effect on reduced crop yields and animals stocks caused to farmers in Western Europe financial losses up to 13 billion Euros (Fink et al., 2004). There is the fear that climate change could increase the stress due to recurrent droughts and have strong negative effects on crops and on the capability to satisfy internal food demand. Heat waves are also perceived to be a major health danger after the 2003 episode that hit Central Europe, heavily affecting north western Mediterranean countries such as Spain, France and Italy. The average summer temperature in Europe exceeded, very likely, the average temperature of any previous summer over the last 500 years (Luterbacher et al., 2004). Unusually high temperature began in June and persisted for the whole July until mid-August. However, from a public health and forest protection perspective, it was the relatively short-lived heatwave that occurred during the first fortnight of August 2003 that had a major impact in Europe (Trigo et al., 2005). The extremely high temperatures in early August were responsible for about 30,000 casualties in Western Europe, half of these in France alone, about 4,200 in Spain, 4,000 in Italy, 2,000 in Portugal and 1,000 in Switzerland. Elderly people, aged over 65 years, children under 4 years, and patients with cardiovascular and chronic respiratory diseases were the most vulnerable categories during such events (Dı´ az et al., 2005). Sea level rise (SLR) represents another potential threat, though the projected SLR, accounting for several possible scenarios, is from 0.09 to 0.88 m by 2100 and should affect only restricted coastal zones. Moreover, also local soil subsidence contribute to increasing the vulnerability of parts of the Mediterranean coast. The most vulnerable situation is in Egypt, where it has been evaluated that a 150 cm sea level rise would imply the loss of 20% of agricultural areas, because of the low lying Nile Delta. Other situations at risk are the Camargue in France and the Venice Lagoon in Italy, where a critical area in the central part of the town is at a level between 70 and 80 cm above the mean sea level (MSL), with a tidal range of about 50 cm.
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
17
Extreme adverse weather events have a high social and economic impact in the Mediterranean area. Especially heavy rainfall and floods, because of many steep river basins which are present in a densely populated territory, are a major concern. On the basis of 166 cases of heavy rainfall and floods and 104 cases of strong winds during 10 years of data, total number of victims above 1,900 and economic losses over 6,000MEuro have been estimated. This figure is likely to be an undervaluation. For Spain alone, and only in four years (1996–1999), the Programme of Natural Hazards of the Spanish Directorate of Civil Defence has accounted 155 deaths by heavy rain and flood events and 28 deaths by storms and strong winds (Jansa et al., 2001a). In Greece, according to the data published from the Hellenic Agricultural Insurance Organization (ELGA) for the year 2002 alone the economic losses from heavy precipitation, hail, floods and extreme winds were over 180MEuro. Single disastrous events have been recorded, such as the 4th November 1966 storm which hit central and north eastern Italy, causing more than 50 deaths. The widespread, huge damages in the eastern Alps, Florence and Venice have been estimated to be more than 1MEuro present-day (De Zolt et al., 2006). Several studies show that the overall cyclonic activity in the Mediterranean region is expected to decrease as an effect of climate change in this century, but extremes could behave differently. The frequency of intense precipitations have been observed to increase in the second half of the 20th century in some regions (e.g. Alpert et al., 2002; Kostopoulou and Jones, 2005) and some simulations show that these events will be more intense in a future emission scenario (Giorgi et al., 2004). This 4th section shows that there are reasons for concern as climate change in the Mediterranean region could affect economically relevant activities and determine social problems related to health and wealth of people. Different level of services, of readiness to emergencies, technological and economic resources, are likely to result in very different adaptation capabilities to environmental changes and new problems. The different economic situations and demographic trends are likely to produce contrasts and conflicts in a condition of limited available resources and environmental stress. Besides description of adaptation strategies, and evaluation of costs, which are needed to reduce such risks, research for understanding of climate mechanisms and the best possible prediction of future climate scenarios has to be continued in order to provide the most reliable information on the evolution of the Mediterranean climate.
Acknowledgements The authors are grateful to E. Elvini for her work on the figures.
18 Mediterranean Climate Variability
References Alpert, P., Neeman, B. U., & Shay-El, Y. (1990). Climatological analysis of Mediterranean cyclones using ECMWF data. Tellus, 42A, 65–77. Alpert, P., Ben-Gai, T., Baharad, A., Benjamini, Y., Yekutieli, D., Colacino, M., Diodato, L., Ramis, C., Homar, V., Romero, R., Michaelides, S., & Manes (2002). The paradoxical increase of Mediterranean extreme daily rainfall in spite of decrease in total values. Geophys. Res. Lett., 29(11), 31-1–31-4. Artale, V., Calmanti, S., & Sutera, A. (2002). North Atlantic THC sensitivity to intermediate level anomalies. Tellus, 54A, 159–174. Ashkenazy, Y., & Stone, P. H. (2003). Box modelling of the Eastern Mediterranean, Submitted to J. Phys. Oceanogr. Barriendos, M., Martin-Vide, J., Pena, J. C., & Rodriguez, R. (2002). Daily meteorological observations in Cadiz – San Fernando. Analysis of the documentary sources and the instrumental data content (1786–1996). Clim. Change, 53, 151–170. Bethoux, J. P., Gentili, B., Raunet, J., & Tailliez, D. (1990). Warming trend in the Western Mediterranean deep water. Nature, 347, 660–662. Bethoux, J. P., Gentili, B., & Tailliez, D. (1998). Warming and freshwater budget change in the Mediterranean since the 1940s, their possible relation to the greenhouse effect. Geophys. Res. Let., 25, 1023–1026. Bethoux, J. P., Gentili, B., Morin, P., Nicolas, E., Pierre, C., & Ruiz-Pino, D. (1999). The Mediterranean Sea: a miniature ocean for climatic and environmental studies and a key for climatic functioning of the North Atlantic. Progress in Oceanography, 44, 131–146. Bolle, H.-J. (Ed) (2003). Mediterranean Climate – Variability and Trends. Springer Verlag, Berlin, Heidelberg, New York. Boscolo, R., & Bryden, H. (2001). Causes of long-term changes in the Aegean Sea deep water. Oceanologica Acta, 24, 519–527. Brunet, M., Aguilar, E., Saladie, O., Sigro´, J., & Lo´pez, D. (2001). The Spanish diurnal warming: a different pattern to the observed on a global scale. Geophysical Research Abstracts, 3, 5332. Buffoni, L., Maugeri, M., & Nanni, T. (1999). Precipitation in Italy from 1833 to 1996. Theor. Appl. Climatol., 63, 33–40. Camuffo, D. (1987). Freezing of the Venetian lagoon since the 9th century AD in comparison to the climate of Western Europe and England. Clim. Change, 10, 43–66. Camuffo, D. (1993). Analysis of the sea surges at Venice from AD 782 to 1990. Theor. Appl. Climatol., 47, 1–14. Camuffo, D. (2002). History of the long series of the air temperature in Padova (1725-today). Clim. Change, 53, 7. Christensen, J. H., Machenhauer, B., Jones, R. G., Sha¨r, C., Ruti, P. M., Castro, M., & Visconti, G. (1997). Validation of present-day regional climate simulations over Europe: LAM simulations with observed boundary conditions. Clim. Dyn., 13, 489. Corte-Real, J., Zhang, X., & Wang, X. (1995). Large-scale circulation regimes and surface climatic anomalies over the Mediterranean. Int. J. Climatol., 15, 1135–1150. Cubasch, U., Meehl, G. A., Boer, G. J., Stouffer, R. J., Dix, M., Noda, A., Senior, C. A., Raper, S., & Yap, K. S. (2001). Projections of future climate change. In: J. T. Houghton, Yihui Ding, and M. Noguer (Eds), Climate Change 2001: The Scientific Basis, (Cambridge University Press, pp. 525–582).
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
19
Dai, A., Fung, I. Y., & DelGenio, A. D. (1997). Surface observed global land precipitation variations during 1900–88. J. Clim., 10, 2943–2962. De´que´, M., & Piedelievre, J. P. (1995). High resolution climate simulation over Europe. Clim. Dyn., 11, 321–339. De´que´, M., Marquet, P., & Jones, R. G. (1998). Simulation of climate change over Europe using a global variable resolution General Circulation Model. Clim. Dyn., 14, 173–189. De Zolt, S., Lionello, P., Malguzzi, P., Nuhu, A., & Tomasin, A. (2006). The disastrous storm of 4th November 1966 on Italy, in preparation. Dı´ az, J., Garcı´ a-Herrera, R., Trigo, R. M., Linares, C., Valente, A., & Hernade´z, E. (2004). The impact of summer 2003 heat wave in Iberia: how should we measure it? International Journal of Biometeorology, doi: 10.1007/s00484-005-0005-8. Du¨menil-Gates, L., & Liess, S. (2001). Impacts of deforestation and afforestation in the Mediterranean region as simulated by the MPI atmospheric GCM. Global and Planetary Change, 30, 309–328. Du¨nkeloh, A., & Jacobeit, J. (2003). Circulation dynamics of Mediterranean precipitation variability 1948–98. Int. J. Climatol., 23, 1843–1866. EMA Eau Mediterraneenne en Atlantique (2004). Study of Mediterranean, water outflow and dispersion in the North-East Atlantic Ocean, V. Thierry, P. Lherminier, & X. Carton, Project Report, 30 pp. Felis, T., Pa¨tzold, J., Loya, Y., Fine, M., Nawar, A. H., & Wefer, G. (2000). A coral oxygen isotope record from the northern Red Sea documenting NAO, ENSO, and North Pacific teleconnections on Middle East climate variability since the year 1750. Paleoceanography, 15, 679–694. Fernandez, J., Saez, J., & Zorita, E. (2003). Analysis of wintertime atmospheric moisture transport and its variability over the Mediterranean basin in the NCEP-Reanalyses. Clim. Res., 23, 195–215. Fink, A. H., Bru¨cher, T., Kru¨ger, A., Leckebusch, G. C., Pinto, J. G., & Ulbrich, U. (2004). The 2003 European summer heatwaves and drought – synoptic diagnosis and impacts. Weather, 59, 209–216. Folland, C. K., et al. (2001). Chapter 2: Observed climate variability and change. In: J.T. Houghton et al. (Eds), Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change (Cambridge University Press, Cambridge, UK and New York, NY, USA), p. 881. Frı´ as, M. D., Ferna´ndez, J., Sa´enz, J., & Rodrı´ guez-Puebla, C. (2005). Operational predictability of monthly average maximum temperature over the Iberian Peninsula using DEMETER simulations and downscaling, Tellus A, (in press). Gala´n, E., Can˜ada, R., Ferna´ndez, F., & Cervera, B. (2001). Annual temperature evolution in the Southern Plateau of Spain from the construction of regional climatic time serie. In: M. Brunet, and Lo´pez (Eds), Detecting and Modelling Regional Climate Change, (Springer Verlag, Berlin, Heidelberg, New York, pp. 119–131). Gibelin, A. L., & De´que´, M. (2003). Anthropogenic climate change over the Mediterranean region simulated by a global variable resolution model. Clim. Dyn., 20, 327–339, doi: 10.1007/s00382-002–0277-1. Giorgi, F. (2002a). Variability and trends of sub-continental scale surface climate in the twentieth century. Part I: observations. Clim. Dyn., 18, 675–691.
20 Mediterranean Climate Variability
Giorgi, F. (2002b). Variability and trends of sub-continental scale surface climate in the twentieth century. Part II: AOGCM simulations. Clim. Dyn., 18, 693–708. Giorgi, F., & Francisco, R. (2000a). Evaluating uncertainties in the prediction of regional climate change. Geophys. Res. Lett., 27, 1295–1298. Giorgi, F., & Francisco, R. (2000b). Uncertainties in regional climate prediction: a regional analysis of ensemble simulations with the HADCM2 coupled AOGCM. Clim. Dyn., 16, 169–182. Giorgi, F., & Mearns, L. O. (1991). Approaches to regional climate change simulations: a review. Rev. Geophys., 29, 191–216. Giorgi, F., Marinucci, M. R., & Visconti, G. (1992). A 2CO2 climate change scenario over Europe generated using a limited area model nested in a General Circulation Model. II: climate change scenario. J. Geophys. Res., 97, 10011–10028. Giorgi, F., Bi, X., & Pal, J. (2004). Mean interannual variability and trends in a regional climate change experiment over Europe. II climate change scenarios (2071–2100). Climate Dynamics, 23, 839–858. Giorgi, F., Hewitson, B., Christensen, J. H., Hulme, M., von Storch, H., Whetton, P., Jones, R. G., Mearns, L. O., & Fu, C. (2001). Regional climate information – evaluation and projections. In: J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, X. Dai, K. Maskell, and C. A. Johnson (Eds), Climate Change 2001: The Scientific Basis. Contribution of Working Group I to Third Assessment Report of the Intergovernmental Panel on Climate Change, (Cambridge University Press, Cambridge, UK and New York, NY, USA, pp. 881). Gonza´lez-Rouco, J. F., Heyen, H., Zorita, E., & Valero, F. (2000). Agreement between observed rainfall trends and climate change simulations in Southern Europe. J. Climate, 13, 3057–3065. Guiot, J., Nicault, A., Rathgeber, C., Edouard, J. L., Guibal, F., Pichard, G., & Till, C. (2005). The last millennium summer temperature variations in Western Europe as based on proxy data. The Holocene, 15, 489, doi: 10.1191/0959683605hl819rp. Hasanean, H. M. (2004). Wintertime surface temperature in Egypt in relation to the associated atmospheric circulation. Int. J. Climatol., 24, 985–999. Hoerling, M. P., Hurrell, J. W., Xu, T., Bates, G. T., & Phillips, A. S. (2004). Twentieth century North Atlantic climate change. Part II: understanding the effect of Indian Ocean warming. Clim. Dyn., 23, 391–405. Hurrell, J. W. (1995). Decadal trends in the North-Atlantic Oscillation-regional temperatures and precipitation. Science, 269, 676–679. Hurrell, J. W., Hoerling, M. P., Phillips, A. S., & Xu, T. (2004). Twentieth century North Atlantic climate change. Part I: assessing determinism. Clim. Dyn., 23, 371–389. Jansa A., Alpert, P., Buzzi, A., & Arbogast, P. (2001a). MEDEX, cyclones that produce high impact weather in the Mediterranean, available at http://medex.inm.uib.es. Josey, S. A. (2003). Changes in the heat and freshwater forcing of the Eastern Mediterranean and their influence on deep water formation. J. Geophys. Res., 108, 3237, doi: 10.1029/2003JC001778. Jung, T., Ferranti, L., & Tompkins, A. M. (2005). Response to the summer 2003 Mediterranean SST anomalies over Europe and Africa. J. Climate, submitted. Kattenberg, A., Giorgi, F., Grassl, H., Meehl, G. A., Mitchell, J. F. B., Stouffer, R. J., Tokioka, T., Weaver, A. J., & Wigley, T. M. L. (1996). Climate models – projections of future climate, in climate change 199. In: J. T. Houghton, L. G. M. Filho, B. A.
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
21
Callander, N. Harris, A. Kattenberg, and K. Maskell (Eds), The Science of Climate Change, (Cambridge University Press, Cambridge, UK, pp. 285–357). Ko¨ppen, W. (1936). Das geographische System der Klimate. In: (Ko¨ppen, W. & R. Geiger) (Eds). Handbuch der Klimatologie 3. (Gebrueder Borntraeger, Berlin, 46 pages). Kostopoulou, E., & Jones, P. D. (2005). Assessment of climate extremes in the Eastern Mediterranean. Meteorol. Atmos. Phys., 89, 69–85. Krichak, S. O., & Alpert, P. (2005a). Decadal trends in the East Atlantic/West Russia pattern and the Mediterranean precipitation. Int. J. Climatol., 25, 183–192. Krichak, S. O., & Alpert, P. (2005b). Signatures of the NAO in the atmospheric circulation during wet winter months over the Mediterranean region. Theor. Appl. Climatol., 82, 27–39. Krichak, S. O., Kishcha, P., & Alpert, P. (2002). Decadal trends of main Eurasian oscillations and the Eastern Mediterranean precipitation. Theor. Appl. Climatol., 72, 209–220. Li, Z. X., (2005). Atmospheric GCM response to an idealized anomaly of the Mediterranean Sea surface temperature, Clim. Dyn., submitted. Lionello, P. (2005). Extreme surges in the Gulf of Venice. present and future climate. In: C. Fletcher, and T. Spencer (Eds), Venice and Its Lagoon, State of Knowledge, (University Press, Cambridge, UK, pp. 59–65). Lionello, P., Dalan, F., & Elvini, E. (2002). Cyclones in the Mediterranean region: the present and the doubled CO2 climate scenarios. Clim. Res., 22, 147–159. Lionello, P., Elvini, E., & Nizzero, A. (2003). Ocean waves and storm surges in the Adriatic Sea: intercomparison between the present and doubled CO2 climate scenarios. Clim. Research, 23, 217–231. Lozier, M. S., Owens, W. B., & Curry, R. G. (1995). The climatology of the North Atlantic. Progr. In Oceanogr., 36, 1–44. Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M., & Wanner, H. (2004). European seasonal and annual temperature variability, trends, and extremes since 1500. Science, 303, 1499–1503, doi: 10.1126/science.1093877. Machenhauer, B., Windelband, M., Botzet, M., Jones, R. G., & De´que´, M., (1996). Validation of present-day regional climate simulations over Europe: Nested LAM and variable resolution global model simulations with observed or mixed layer ocean boundary conditions, Max Planck Institut fu¨r Meteorologie, rep. 191, 52 pp. Machenhauer B., Windelband, M., Botzet, M., Christensen, J. H., De`que`t, M., Jones, R. G., Ruti, P. M., & Visconti, G., (1998). Validation and analysis of present-day climate and climate change simulations over Europe, Max Planck Institut fu¨r Meteorologie, rep.275, 87 pp. Malanotte-Rizzoli, P., Manca, B., Ribera d’Alcala`, M., Teocharis, A., Brenner, S., Budillon, G., & Oszoy, E. (1999). The Eastern Mediterranean in the 80’s and in the 90’s: the big transition in the intermediate and deep circulations. Dyn. Atmos. Oceans, 29, 365–395. Manca, B. B., Budillon, G., Scarazzato, P., & Ursella, L. (2003). Evolution of dynamics in the Eastern Mediterranean affecting water mass structures and properties in the Ionian and Adriatic Seas. J. Geophys. Res., 108, 81029, doi:1029/2002JC001664. Mann, M. E. (2002). Large-scale climate variability and connections with the Middle East in past centuries. Clim. Change, 55, 287–314.
22 Mediterranean Climate Variability
Marinucci, M. R., & Giorgi, F. (1992). A 2CO2 climate change scenario over Europe generated using a limited area model using nested in a general circulation model. I: present day simulation. J. Geophys. Res., 97, 9989–10009. Mariotti, A., Zeng, N., & Lau, K. M. (2002a). Euro-Mediterranean rainfall and ENSO – a seasonally varying relationship. Geophys. Res. Lett., 29, 1621. Mariotti, A., Struglia, M. V., Zeng, N., & Lau, K.-M. (2002b). The hydrological cycle in the Mediterranean region and implications for the water budget of the Mediterranean Sea. J. Climate, 15, 1674–1690. Maugeri, M., Buffoni, L., & Chlistovsky, F. (2002). Daily Milan temperature and pressure series (1763–1998): history of the observations and data and metadata recovery. Clim. Change, 53, 101–117. May, W., & Bengtsson, L. (1998). The signature of ENSO in the Northern Hemisphere midlatitude seasonal mean flow and high-frequency intraseasonal variability. Meteorol. Atmos. Phys., 69, 81–100. Millot, C. (1999). Review paper circulation in the Western Mediterranean sea. Journal of Marine System, 20, 423–442. Mun˜oz-Dı´ az, D., & Rodrigo, F. S. (2004). Impacts of the North Atlantic oscillation on the probability of dry and wet winters in Spain. Climate Research, 27, 33–43. New, M. G., Hulme, M., & Jones, P. D. (2000). Representing twentieth-century spacetime climate variability. Part II: development of 1901–1996 monthly grids of terrestrial surface climate. J. Climate, 13, 2217–2238. New, M., Todd, M., Hulme, M., & Jones, P. D. (2001). Precipitation measurements and trends in the twentieth century. Int. J. Climatol., 21, 1899–1922. Nittis, K., Lascaratos, A., & Theocharis, A. (2003). Dense water formation in the Aegean Sea: numerical simulations during the Eastern Mediterranean transient. J. Geophys. Res., 108, 8120, doi: 10.1029/2002JC001352. Ovchinnikov, I. M. (1966). Circulation in the surface and intermediate layer of the mediterranean. Oceanology, 24, 168–173. Parry, M. L. (Ed). The European ACACIA Project. Jackson Environment Institute, University of Norwich, Anglia, UK, 320 pp. Pauling, A., Luterbacher, J., Casty, C., & Wannner, H. (2005). 500 years of gridded highresolution precipitation reconstructions over Europe and the connection to large-scale circulation. Clim. Dyn., accepted. Pinardi, N., & Masetti, E. (2000). Variability of the large scale general circulation of the Mediterranean Sea from observations and modelling: a review. Palaeogeography, palaeoclimatology, palaeoecology, 158, 153–174. Plakhin, Ye A. (1972) Vertical winter circulation in the Mediterranean. Oceanology, 12(3), 344–351. POEM Group (1992). The general circulation of the Eastern Mediterranean. Earth-Science Reviews, 32, 285–309. Price, C., Stone, L., Rajagopalan, B., & Alpert, P. (1998). A possible link between El Nino and precipitation in Israel. Geophys. Res. Lett., 25, 3963–3966. Potter, R. A., & Lozier, M. S. (2004). On the warming and salinification of the Mediterranean outflow waters in the North Atlantic. Geophys. Res. Lett., 31, L01202, doi:10.1029/2003GL018161. Pozo-Va´zquez, D., Esteban-Parra, M. J., Rodrigo, F. S., & Castro-Diez, Y. (2001). A study of NAO variability and its possible non-linear influences on European surface temperature. Clim. Dyn., 17, 701–715.
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
23
Price, C., Stone, L., Rajagopalan, B., & Alpert, P. (1998). A possible link between El Nino and precipitation in Israel. Geophys. Res. Lett., 25, 3963–3966. Quadrelli, R., Pavan, V., & Molteni, F. (2001). Wintertime variability of Mediterranean precipitation and its links with large-scale circulation anomalies. Clim. Dyn., 17, 457–466. Rahmstorf, S. (1998). Influence of Mediterranean Outflow on climate. Eos., 79, 281–282. Raisa¨nen, U., Hansson, J., Ullerstig, U., Doscher, A., Graham, R., Jones, L. P., Meier, C., Samuelsson, H. E. M., & Willen, P. (2004). European climate in the late twenty-first century: regional simulations with two driving global models and two forcing scenarios. Clim. Dyn., 22, 13–32. Reale, O., & Dirmeyer, P. (2000). Modeling the effects of vegetation on Mediterranean climate during the Roman Classical Period – Part I: climate history and model sensitivity. Global Planet, 25, 163–184. Reale, O., & Shukla, J. (2000). Modeling the effects of vegetation on Mediterranean climate during the Roman Classical Period: Part II: model simulation. Global Planet, 25, 185–214. Reid, J. L. (1994). On the total geostrophic circulation of the North Atlantic Ocean: flow patterns, tracers and transports. Progr. in Oceanogr., 33, 1–92. Rodo´, X. (2001). Reversal of three global atmospheric fields linking changes in SST anomalies in the Pacific, Atlantic and Indian Oceans at tropical latitudes and midlatitudes. Clim. Dynamics, 18, 203–217. Rodo´, X., Baert, E., & 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. Clim. Din., 13(4), 275–284. Rodrigo, F. S. (2002). Changes in climate variability and seasonal rainfall extremes: a case study from San Fernando (Spain), 1821–2000. Theor. Appl. Climatol., 72, 193–207. Rodwell, M. J., & Hoskins, B. J. (1996). Monsoons and the dynamic of deserts. Quar. J. Roy. Meteorol. Soc., 122, 1385–1404. Rogers, J. C. (1990). Patterns of low-frequency monthly sea level pressure variability (1899–1986) and associated wave cyclone frequencies. J. Climate, 3, 1364–1379. Rogers, J. C. (1997). North Atlantic storm track variability and its association to the North Atlantic Oscillation and climate variability of the Northern Europe. J. Climate, 10, 1635–1647. Roether, W., & Schlitzer, R. (1991). Eastern Mediterranean deep water renewal on the basis of chlorofluoromethane and tritium data. Dyn. Atmos. and Oceans, 15, 333–354. Roether, W., Schlosser, P., Kuntz, R., & Weis, W.,(1983). Transient tracers studies of the thermohaline circulation of the Mediterranean. In: H. Charnock (Ed.), MS for Mediterranean Circulation, NATO Workshop, 44 pp. Roether, W., Manca, B. B., Klein, B., Bregant, D., Georgopoulos, D., Beitzel, V., Kovacevic, V., & Luchetta, A. (1996). Recent changes in Eastern Mediterranean deep waters. Science, 271, 333–335. Rohling, E. J., & Bryden, H. L. (1992). Man-induced salinity and temperature increases in Western Mediterranean deep water. J. Geophys. Res., 97, 11191–11198. Ro¨tter, R., & van de Geijn, S. C. (1999). Climate change effects on plant growth, crop yield and livestock. Clim. Change, 43, 651–681.
24 Mediterranean Climate Variability
Rowell, D. P. (2003). The impact of Mediterranean SSTs on the Sahelian rainfall season. J. Climate, 16, 849–862. Sa´enz, J., Rodrı´ guez-Puebla, C., Ferna´ndez, J., & Zubillaga, J. (2001). Interpretation of interannual winter temperature variations over Southwestern Europe. J. Geophys. Res., 106, 20641–20651. Semazzi, F. H., & Sun, L. (1997). The role of orography in determining the Sahelian Climate. Int. J. Climatol., 17, 581–596. Samuel, S., Haines, K., Josey, S., & 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. J. Geophys. Res., 104(C4), 7771–7784. Shindell, D. T., Schmidt, G. A., Miller, R. L., & Mann, M. E. (2003). Volcanic and solar forcing of climate change during the preindustrial era. J. Climate, 16, 4094–4107, doi:10.1175/1520–0442. Skliris, N., & Lascaratos, A., (2004). Impacts of the Nile River damming on the thermohaline circulation and water mass characteristics of the Mediterranean Sea. J. Marine Systems, in press. Somot, S. (2005). Mode´lisation climatique du bassin me´diterrane´en: variabilite´ et sce´narios de changement climatique. PhD Thesis, Universite´ Paul Sabatier, Toulouse, France. Tarawneh, Q., & Kadioglu, M. (2003). An analysis of precipitation climatology in Jordan. Theor. Appl. Climatol., 74, 123–136. Theocharis, A., & Georgopoulos, D. (1993). Dense water formation over the Samothraki and Lemnos plateaux in the North Aegean Sea (Eastern Mediterranean Sea). Continental Shelf Research, 13, 919–939. Theocharis, A., Klein, B., Nittis, K., & Roether, W. (2002). Evolution and status of the Eastern Mediterranean Transient (1997–1999). J. Mar. Syst., 33–34, 91–116. Till, C., & Guiot, J. (1990). Reconstruction of precipitation in Morocco since 1100 AD based on Cedrus atlantica tree-ring widths. Quaternary Res., 33, 337–351. Touchan, R., Garfin, G. M., Meko, D. M., Funkhouser, G., Erkan, N., Hughes, M. K., & Wallin, B. S. (2003). Preliminary reconstructions of spring precipitation in southwestern Turkey from tree-ring width. Int. J. Climatol., 23, 157–171. Touchan, R., Xoplaki, E., Funkhouser, G., Luterbacher, J., Hughes, M. K., Erkan, N., Akkemik, U¨. & Stephan, J. (2005). Reconstructions of spring/summer precipitation for the Eastern Mediterranean from tree-ring widths and its connection to large-scale atmospheric circulation, Clim. Dyn., doi: 10.1007/s00382-005–0016-5. Trigo, I. F., Davies, T. D., & Bigg, G. R. (1999). Objective climatology of cyclones in the Mediterranean region. J. Clim., 12, 1685–1696. Trigo, R. M., & DaCamara, C. (2000). Circulation weather types and their impact on the precipitation regime in Portugal. International Journal of Climatology, 20, 1559–1581. Trigo, I. F., Trevor, T. D., & Bigg, G. R. (2000). Decline in Mediterranean rainfall caused by weakening of Mediterranean cyclones. Geophys. Res. Lett., 27, 2913–2916. Trigo, R. M., Osborn, T. J., & Corte-Real, J. (2002). The North Atlantic Oscillation influence on Europe: climate impacts and associated physical mechanisms. Clim. Res., 20, 9–17 10. Trigo, R. M., Pozo-Vazquez, D., Osborn, T. J, Castro-Diez, Y., Ga´mis-Fortis, S., & Esteban-Parra, M. J. (2004). North Atlantic Oscillation influence on precipitation, river flow and water resources in the Iberian Peninsula. Int. J. Climatol., 24, 925–944.
The Mediterranean Climate: An Overview of the Main Characteristics and Issues
25
Trigo, R. M., Garcı´ a-Herrera, R., Dı´ az, J., Trigo, I. F., & Valente, A. (2005). How exceptional was the early August 2003 heatwave in France. Geophys. Res. Lett., 32, L10701, doi:10.1029/2005GL022410. Tsimplis, M. N., & Baker, T. F. (2000). Sea level drop in the Mediterranean Sea: an indicator of deep water salinity and temperature changes. Geophys. Res. Lett., 27, 1731–1734. Tsiourtis, N. X. (2001). Drought management plans for the Mediterranean region. Report of the water development department. Nicosia, Cyprus. Available through: http://www.pio.gov.cy/wdd/eng/scientific_articles/archieve2001/article03.htm. Ulbrich, U., & Christoph, M. (1999). A shift in the NAO and increasing storm track activity over Europe due to anthropogenic greenhouse gas. Clim. Dyn., 15, 551–559. Yakir, D., Lev-Yadun, S., & Zangvil, A. (1996). El Nino and tree growth near Jerusalem over the last 20 years. Global Change Biology, 2, 101–105. Xoplaki, E. (2002) Climate variability over the Mediterranean. PhD Thesis, University of Bern, Switzerland. Available through: http://sinus.unibe.ch/klimet/ docs/phd_xoplaki.pdf. Xoplaki, E., Gonzalez-Rouco, F. J., Luterbacher, J., & Wanner, H. (2003). Mediterranean summer air temperature variability and its connection to the large-scale atmospheric circulation and SSTs. Clim. Dyn., 20, 723–739, doi: 10.1007/ s00382-003–0304-x. Xoplaki, E., Gonzalez-Rouco, F. J., Luterbacher, J., & Wanner, H. (2004). Wet season Mediterranean precipitation variability: influence of large-scale dynamics. Clim. Dyn., 23, 63–78, doi: 10.1007/s00382-004–0422-0. Xoplaki, E., Luterbacher, J., Paeth, H., Dietrich, D., Steiner, N., Grosjean, M., & Wanner, H. (2005). European spring and autumn temperature variability and change of extremes over the last half millennium. Geophys. Res. Lett., 32, L15713, doi: 10.1029/ 2005GL023424. Zervakis, V., Georgopoulos, D., & Drakopoulos, P. G. (2000). The role of the north Aegean in triggering the recent Eastern Mediterranean climatic changes. J. Geophys. Res., 105, 26103–26116. Ziv, B., Saaroni, H., & Alpert, P. (2004). The factors governing the summer regime of the Eastern Mediterranean. Int. J. Clim., 24, 1859–1871. Zorita, E., von Storch, H., Gonzalez-Rouco, F. J., Cubasch, U., Luterbacher, J., Fischer-Bruns, I, Legutke, S., & Schlese, U. (2004). Climate evolution in the last five centuries simulated by an atmosphere-ocean model: global temperatures, the North Atlantic Oscillation and the Late Maunder Minimum. Meteorol. Zeitschrift, 13, 271–289.
Notes Documentation on water resources, demographic trends, economics and climate change in the Mediterranean are available at the Center of studies ‘‘Le plan Bleu’’ http://www.planbleu.org/. This documents used material from the following reports: Demography in the Mediterranean region: situation and projections ATTANE´, Isabelle; COURBAGE, Youssef, Plan Bleu, 2004
26 Mediterranean Climate Variability
L’eau des Me´diterrane´ens: situation et perspectives MARGAT Jean, avec la collaboration de Se´bastien TREYER, PNUE. PAM. Plan Bleu, 2004 The present status of knowledge on global climatic change; its regional aspects and impacts in the Mediterranean region MEDIAS-FRANCE PNUE. PAM. Plan Bleu, 2001
Chapter 1
Mediterranean Climate Variability over the Last Centuries: A Review Ju¨rg Luterbacher,1 Elena Xoplaki,1 Carlo Casty,1 Heinz Wanner,1 Andreas Pauling,2 Marcel Ku¨ttel,2 This Rutishauser,2 Stefan Bro¨nnimann,3 Erich Fischer,3 Dominik Fleitmann,4 Fidel J. Gonza´lez-Rouco,5 Ricardo Garcı´ a-Herrera,5 Mariano Barriendos,6 Fernando Rodrigo,7 Jose Carlos Gonzalez-Hidalgo,8 Miguel Angel Saz,8 Luis Gimeno,9 Pedro Ribera,10 Manola Brunet,11 Heiko Paeth,12 Norel Rimbu,13 Thomas Felis,14 Jucundus Jacobeit,15 Armin Du¨nkeloh,16 Eduardo Zorita,17 Joel Guiot,18 Murat Tu¨rkes,19 Maria Joao Alcoforado,20 Ricardo Trigo,21 Dennis Wheeler,22 Simon Tett,23 Michael E. Mann,24 Ramzi Touchan,25 Drew T. Shindell,26 Sergio Silenzi,27 Paolo Montagna,27 Dario Camuffo,28 Annarita Mariotti,29 Teresa Nanni,30 Michele Brunetti,30 Maurizio Maugeri,31 Christos Zerefos,32 Simona De Zolt,33 Piero Lionello,33 M. Fatima Nunes,34 Volker Rath,35 Hugo Beltrami,36 Emmanuel Garnier37 and Emmanuel Le Roy Ladurie38 1
NCCR Climate and Institute of Geography, University of Bern, Bern, Switzerland (
[email protected],
[email protected],
[email protected],
[email protected]) 2 Institute of Geography, University of Bern, Bern, Switzerland (
[email protected],
[email protected],
[email protected]) 3 ETH Zurich, Institute for Atmospheric and Climate Science, Zurich, Switzerland (
[email protected],
[email protected]) 4 Institute of Geological Sciences, University of Bern, Bern, Switzerland (
[email protected]) 5 Department of Physics, University Complutense, Madrid, Spain (
[email protected],
[email protected]) 6 Department of Modern History, University of Barcelona, Barcelona, Spain (
[email protected]) 7 Department of Applied Physics, University of Almeria, Spain (
[email protected]) 8 Department of Geography, University of Zaragoza, Spain (
[email protected],
[email protected]) 9 Facultad de Ciencias de Ourense, University of Vigo, Orense, Spain (
[email protected]) 10 Depto. CC. Ambientales, University Pablo de Olavide, Sevilla, Spain (
[email protected])
28 Mediterranean Climate Variability
11
Climate Change Research Group, University Rovira i Virgili, Tarragona, Spain (
[email protected]) 12 Institute of Meteorology, University of Bonn, Bonn, Germany (
[email protected]) 13 Department of Geosciences, Bucharest University, Bucharest-Magurele, Romania, and Alfred-Wegener-Institute for Polar and Marine Research, Bremerhaven, Germany (
[email protected]) 14 DFG-Research Center for Ocean Margins, University of Bremen, Bremen, Germany (
[email protected]) 15 Institute of Geography, University of Augsburg, Augsburg, Germany (
[email protected]) 16 Institute of Geography, University of Wu¨rzburg, Germany (
[email protected]) 17 Department of Paleoclimate, Institute for Coastal Research, GKSS, Geesthacht, Germany (
[email protected]) 18 CEREGE CNRS UMR 6635, BP 80, Aix-en-Provence cedex 4, France (
[email protected]) 19 Department of Geography, Canakkale Onsekiz Mart University, Canakkale, Turkey (
[email protected]) 20 Centro de Estudos Geogra´ficos. Universidade de Lisboa, FLUL, 1600-214 Lisbon, Portugal (
[email protected]) 21 Faculdade de Cieˆncias, Departamento de Fı´sica, Universidade de Lisboa, CGUL (
[email protected]) 22 University of Sunderland, U.K. (
[email protected]) 23 Met Office, Hadley Centre, Reading RG6 6BB, U.K. (
[email protected]) 24 Department of Meteorology and Earth & Environmental Systems Institute (ISSI), The Pennsylvania State University, University Park, PA 16802, USA (
[email protected]) 25 Laboratory of Tree-Ring Research, The University of Arizona, Tucson, USA (
[email protected]) 26 NASA Goddard Institute for Space Studies and Columbia University, New York, New York, USA (
[email protected]) 27 ICRAM – Central Institute for Marine Research, Rome, Italy (
[email protected],
[email protected]) 28 Institute of Atmospheric Sciences and Climate, National Research Council, Padova, Italy (
[email protected]) 29 ENEA, Italy (
[email protected]) 30 Institute of Atmospheric Sciences and Climate, Bologna, Italy (
[email protected],
[email protected]) 31 Istituto di Fisica Generale Applicata, University of Milan, Milan, Italy (
[email protected]) 32 Faculty of Geology, Laboratory of Climatology and Atmospheric Environment, University of Athens, Athens, Greece (
[email protected])
Mediterranean Climate Variability over the Last Centuries: A Review
29
33
Department of Physics, University of Padua, Padova, Italy (
[email protected],
[email protected]) 34 University of E´vora Portugal, Department of History, E´vora, Portugal (
[email protected]) 35 Applied Geophysics, RWTH Aachen University, Germany (
[email protected]) 36 Environmental Sciences Research Center, St. Francis Xavier University, Antigonish, Nova Scotia, Canada (
[email protected]) 37 CNRS UMR 6583, University of Caen, Laboratory of Climatic and Environmental Sciences CEA CNRS, Gif-sur-Yvette, France (
[email protected]) 38 Colle`ge de France, Paris, France (
[email protected])
1.1. Introduction A necessary task for assessing to which degree the industrial period is unusual against the background of pre-industrial climate variability, is the reconstruction and interpretation of temporal and spatial patterns of climate in earlier centuries. The comparison of past climate reconstructions with numerical models can enhance our dynamical and physical understanding of the relevant processes. As widespread, direct measurements of climate variables are only available about one to two centuries back in time, it is necessary to use indirect indicators or ‘‘proxies’’ of climate variability, which are recorded in natural archives (coral reefs, ice cores, tree rings, boreholes, speleothems, etc.). These archives record, by their biological, chemical and physical nature, climate-related phenomena (Jones and Mann, 2004). The use of natural proxies, especially in a quantitative way, is a more recent tool in paleoclimate research. Additionally, documentary evidence provides information about past climate variability by means of direct or indirect descriptions of climate-related phenomena (e.g. Pfister, 1999, 2005; Bartholy et al., 2004; Chuine et al., 2004; Bra´zdil et al., 2005; Glaser and Stangl, 2005; Guiot et al., 2005; Przybylak et al., 2005; Le Roy Ladurie, 2004, 2005). The Mediterranean area offers a broad spectrum of long high-quality instrumental time series, documentary information and natural archives, both in time and space, making this area is ideal for climate reconstructions of past centuries, as well as for the analysis of changes in climate extremes and socio-economic impacts prior to the instrumental period. Documentary evidence such as written sources, paintings or flood markers have widely been used for climate reconstructions (e.g. Luterbacher et al., 2004;
30 Mediterranean Climate Variability
Casty et al., 2005; Guiot et al., 2005; Xoplaki et al., 2005). In the Mediterranean area, documentary evidence reaches more than two millennia back in time. The question to what extent climate has changed since the classical epoch and whether or not the extensive deforestation was the cause for climatic change in this region has been discussed since the eighteenth century (Bro¨nnimann, 2003). Mann (1790) concluded from written sources that climate has become progressively warmer and drier over time, which he could not explain by land use changes. Others did not support the idea of a progressive climate change. Ideler (1832), for instance, criticized Mann in being too trustful in his documentary sources. The question was discussed by many others, partly in the context of the deforestation and reforestation debate in the nineteenth century (e.g. Rico Sinobas, 1851; Arago, 1858; Fischer, 1879; Gu¨nther, 1886; Bru¨ckner, 1890). Hence, studies on past Mediterranean climate variability as well as the use of documentary evidence for climate reconstruction have a long scientific tradition. There are distinct differences in the temporal resolution among the various proxies. Some of the proxy records are annually or even higher resolved (documentary data, growth and density measurements from tree rings, corals, annually resolved ice cores, laminated ocean and lake sediment cores, and speleothems) and hence record year-by-year patterns of climate in past centuries (Jones and Mann, 2004). Other proxies such as boreholes capture the lowfrequency signal. Jones and Mann (2004) review the strengths of each proxy source with emphasis on the potential weaknesses and caveat. Climate records for land areas (compared to marine records) exhibit a high degree of geographical variability due to local peculiarities. The use of a broad collection of proxies may help in disentangling the geographical complexity (e.g. Pla and Catalan, 2005). Properties such as sensitivity, reproducibility, local availability and continuity through time differ among them (Mann, 2002a; Pauling et al., 2003; Jones and Mann, 2004). A number of previous studies have focused on global to hemispheric temperature reconstructions over the past few centuries to millennia, based on both empirical proxy data (Bradley and Jones, 1993; Jones et al., 1998; Overpeck et al., 1997; Briffa et al., 1998, 2001, 2002, 2004; Mann et al., 1998, 1999; Crowley and Lowery, 2000; Harris and Chapman, 2001; Esper et al., 2002, 2004, 2005a,b; Beltrami and Bourlon, 2004; Cook et al., 2004; Huang, 2004; Pollack and Smerdon, 2004; Moberg et al., 2005) and model simulations including forcing data (e.g. Crowley, 2000; Waple et al., 2002; Bauer et al., 2003; Gerber et al., 2003; Gonza´lez-Rouco et al., 2003a,b, 2006; Rutherford et al., 2003, 2005; von Storch et al., 2004; Zorita et al., 2004, 2005; Goosse et al., 2005a,b,c; Mann et al., 2005; Stendel et al., 2006; van der Schrier and Barkmeijer, 2005). Several of the temperature reconstructions reveal that the late twentieth century warmth is unprecedented at hemispheric scales, and can only be explained by
Mediterranean Climate Variability over the Last Centuries: A Review
31
anthropogenic, greenhouse gas (GHG) forcing (Jones and Mann, 2004 and references therein). Hemispheric temperature reconstructions cannot provide information about regional-scale climate variations. Several sources point to differing courses of temperature change in Europe and the generally greater amplitude of variations than recorded for the Northern Hemisphere (NH) (e.g. Mann et al., 2000; Jones and Mann, 2004; Luterbacher et al., 2004; Bra´zdil et al., 2005; Casty et al., 2005a; Guiot et al., 2005; Xoplaki et al., 2005). For instance, the European heat wave of summer 2003 was a regional expression of an extreme event, much larger in amplitude than extremes at hemispheric scales (Chuine et al., 2004; Luterbacher et al., 2004; Pal et al., 2004; Rebetez, 2004; Scha¨r et al., 2004; Scho¨nwiese et al., 2004; Stott et al., 2004; Menzel, 2005; Trigo et al., 2005; Bu¨ntgen et al., 2005a,b; Casty et al., 2005a). The 2003 June–August mean temperature for the larger Mediterranean land area exceeded the 1961–1990 reference period by around 2.3 C (Luterbacher et al., 2004; Stott et al., 2004), and makes it the warmest summer for more than the last 500 years (Luterbacher et al., 2004). Stott et al. (2004) suggest, that human influence has likely doubled the risk of a heatwave exceeding this threshold magnitude of around 2 C in this area. We would like to point out that this review mainly deals with the climate variability over the last few centuries covering the larger Mediterranean area of 30 N–47 N and 10 W–40 E. We will not report about climate reconstructions, climate change and variability over longer timescales. There are many publications (e.g. Araus et al., 1997; Jalut et al., 1997, 2000; Davis et al., 2003; Rimbu et al., 2003a; Battarbee et al., 2004; Felis et al., 2004) dealing with these topics. The new book edited by Battarbee et al. (2004, and references therein) provides a major synthesis of evidence for past climate variability at the regional- and continental-scale across Europe and Africa, including parts of the Mediterranean. It focuses on two complementary timescales, the Holocene (approximately the last 11,500 years) and the last glacial–interglacial cycle (approximately the last 130,000 years). We first report on the availability and potential of long, homogenized instrumental data, documentary and natural proxies to reconstruct aspects of past climate at local- to regional-scales within the larger Mediterranean area, including climate extremes and the incidence of natural disasters. We then turn to recent attempts of large-scale multi-proxy field reconstructions for the Mediterranean land areas and discuss the importance of natural and documentary proxies for regional seasonal temperature and precipitation reconstructions. In Section 1.4 of this chapter, we analyse the reconstructions with respect to the evolution of the averaged Mediterranean winter temperature and precipitation back to 1500 and discuss uncertainties, trends, cold and warm, wet and dry periods and present climate fields of extremes covering the last centuries.
32 Mediterranean Climate Variability
We also briefly address the question how major tropical volcanos influenced Mediterranean winter climate over the past. We investigate whether particularly dry and wet as well as cold and warm Mediterranean winters occurred more frequently during the twentieth century, when climate is increasingly affected by human activity through emissions of GHGs, than earlier. For sub-areas the Palmer Drought Severity Index (PDSI; e.g. Palmer, 1965; Guiot et al., 2005) is derived and changes are discussed in the context of the past. We then analyse the large-scale atmospheric circulation influence on past and present Mediterranean winter climate inclusive extremes at seasonal scale (Section 1.5). In Section 1.6 we will briefly comment on the possible teleconnections between Mediterranean climate and other parts of the NH related to past climate. The relations between variability in the Mediterranean region and global tropical oceans, El Nin˜o-Southern Oscillation (ENSO), Indian and African Monsoon and the mid-latitudes for the instrumental period will be discussed in Chapters 2 and 3. Finally, in Section 1.7 the climate reconstructions are compared with forced simulations of the climate models ECHO-G and HadCM3. Thereby we address the role of external forcing, including natural (e.g. volcanic and solar irradiance) and anthropogenic (GHG and sulphate aerosol) influences, and natural, internal variability in the coupled ocean–atmosphere system at subcontinental scale. We end with conclusions and an outlook on future directions related to Mediterranean past climate.
1.2. Past Regional Mediterranean Climate Evidence and Extremes 1.2.1. Evidence from Early Instrumental and Documentary Data The Mediterranean offers a few long, high-quality and homogenized instrumental station series covering the past few centuries. In the late 1500 Galileo, the Grand Duke of Tuscany and the Accademia del Cimento invented the modern meteorological instruments (thermometer, barometer, hygrometer) and started regular observations. From Bologna, Padova, Milan and the Po Plain (Italy) there are temperature, precipitation and pressure series available at daily-tomonthly resolution (Table 1; Camuffo, 1984, 2002a,b,c; Brunetti et al., 2001; Cocheo and Camuffo, 2002; Maugeri et al., 2002a,b, 2004). Long instrumental temperature, precipitation and pressure series available from southern and northeastern Spain (Rodriguez et al., 2001; Barriendos et al., 2002; Rodrigo, 2002; Vinther et al., 2003b). Alcoforado et al. (1997, 1999) and Taborda et al. (2004) used combined weather and climate information from Portuguese documentary sources and early instrumental data back to the eighteenth century (Table 1). Apart from natural proxies (Section 1.2.2), documentary proxy evidence
Table 1: Compilation of long early homogenized instrumental data and documentary proxy evidence from the Mediterranean (see Section 1.2.1 for details). Location Padova (Italy)
Time period
Type of ‘‘proxy’’
References
Precipitation
Camuffo (1984)
Temperature, pressure Temperature, precipitation Temperature, pressure Pressure
Camuffo (2002a,b,c); Cocheo & Camuffo (2002) Brunetti et al. (2001)
Maugeri et al. (2003)
Pressure
Rodriguez et al. (2001)
Temperature, pressure
Barriendos et al. (2002)
Monthly
Precipitation
Rodrigo (2002)
Monthly
Pressure
Vinther et al. (2003b)
Monthly, seasonal
Temperature, precipitation, drought Temperature, precipitation
Taborda et al. (2004)
Daily/ monthly Daily/ monthly Daily/ monthly Daily/ monthly Daily/ monthly Daily/ monthly Daily/ monthly
Daily and monthly
Maugeri et al. (2002a,b)
Alcoforado et al. (1997, 1999)
33
(Continued)
Mediterranean Climate Variability over the Last Centuries: A Review
1721–present Instrumental records Padova (Italy) 1721–present Instrumental records Bologna (Italy) 1813–present Instrumental records Milan (Italy) 1763–present Instrumental records Po Plain (Italy) 1765–present Instrumental records Barcelona (Spain) 1780–present Instrumental records 1786–present Documentary and Cadiz-San Fernando (Spain) instrumental records 1821–present Instrumental San Fernando (Spain) records Gibraltar (UK) 1821–present Instrumental records Southern Portugal 1700–1799 Documentary and early instrumental records Lisbon (Portugal) 1815–present Early instrumental
Temporal resolution Parameter
Table 1: Continued. Time period
Type of ‘‘proxy’’
Temporal resolution Parameter
References
Spanish Mediterranean Coast; 37 N–43 N, 2 W–4 E Catalonia (Spain; 40 N–43 N, 0 E–4 E) Catalonia (Spain) 40 N–43 N, 0 E–4 E Barcelona (Spain)
1300–1970
Documentary
Extremes
Barriendos & Martin-Vide (1998)
1521–1825
Documentary, rogations
Extremes, annual
Frequency of extreme events (flood magnitude) Drought indices
1521–1825
Documentary, rogations
Extremes
Frequency of drought, weather
Barriendos & Llasat (2003)
1521–2000
Rogations
Precipitation, droughts, floods
Murcia (Spain)
1570–2000
Rogations
Barriendos (1997); Barriendos & Martin-Vide (1998) Barriendos & Rodrigo (2005)
Spain, 5 points 36 N–44 N; 9 W–4 E Castille (Spain)
1675–1715
Documentary, rogations
Extremes, monthly and seasonal Extremes, monthly and seasonal Annual
1634–1648
Correspondence and documentary evidence
Extremes, irregular
Andalusia (Spain)
1500–1997
Miscellaneous documentary evidence
Extremes, monthly, seasonal
Precipitation, droughts, floods Rainfall indices, droughts, excessive rainfall Precipitation (snowfalls, droughts, floods), temperature, general weather conditions Precipitation, droughts, floods
Martin-Vide & Barriendos (1995)
Barriendos (1997)
Rodrigo et al. (1998)
Rodrigo et al. (1999, 2000)
34 Mediterranean Climate Variability
Location
1595–1836
Central Portugal
1663–1665
Last centuries– present
Southern Portugal 1675–1715
Southeastern France
1200–1999
Agricultural records Miscellaneous documentary, early instrumental data
Extended winter Daily, monthly and seasonal
Correspondence
Irregular (from 2 to 17 monthly weather information) Monthly, seasonal
Different documentary sources Documentary and early instrumental
Wet period
Burgundy (France) 1370–present Grape harvest Spring– from documentary Summer evidence Summer Marseille (France) 1100–1994 Documentary evidence, isotopes, tree rings,
Precipitation, wetness/drought Compilation of different past climate-related topics, weather, temperature and precipitation, logbook information, tropical cyclones, etc. Temperature, precipitation (droughts, floods)
Garcı´ a-Herrera et al. (2003b) http://www.ucm.es/info/ reclido/
Temperature, precipitation (droughts, floods) Hydrological parameters (e.g. floods, annual maxima discharges) Temperature
Alcoforado et al. (2000)
Temperature
Daveau (1997)
Guilbert (1994); Pichard (1999)
Le Roy Ladurie (1983, 2004, 2005); Chuine et al. (2004) Guiot et al. (2005)
35
(Continued)
Mediterranean Climate Variability over the Last Centuries: A Review
Canary Islands (Spain) Spain
Table 1: Continued. Time period
Venice (Italy)
Type of ‘‘proxy’’
Temporal resolution
Parameter
References
6th century– Documentary present sources 1700–present Iconographic sources 782–present Documentary sources
Irregular, winter severity Irregular
Camuffo (1987)
Northern and Central Italy
579–present
Documentary sources
Irregular
Tiber and Po rivers (Italy)
BC 414–present
Documentary sources
Irregular
Padova (Italy)
1300–present Documentary sources, early instrumental records Last Documentary millennium sources
Irregular
Frozen lagoon, ‘‘temperature’’ Algae belt level and sea level rise Flooding tides, Scirocco and Bora wind Locust invasions, Scirocco and Bora wind Precipitation, flood magnitude/ frequency Storms, hailstorm, thunderstorm
Camuffo et al. (2000a)
1374–present Documentary sources
Irregular
Atmospheric circulation, sea storms High-pressure situations (based on dry fog and volcanic clouds)
Venice (Italy) Venice (Italy)
Western Mediterranean and Adriatic sea Italy
Irregular
Irregular
Camuffo & Sturaro (2003, 2004) Camuffo (1993); Enzi & Camuffo (1995) Camuffo & Enzi (1991)
Camuffo & Enzi (1996)
Camuffo et al. (2000b)
Camuffo & Enzi (1994b, 1995c)
36 Mediterranean Climate Variability
Location
Irregular, extremes
1500–1799
Southern Italy
1300–1900
Irregular
Sicily (Italy)
1565–1915
Religious processions 1580–present Documentary evidence and instrumental data 1200–1900 Documentary evidence 1675–1830 Documentary evidence
Extremes
Egypt
622–present
Nilometer, documentary evidence
Irregular, seasonal
Eastern Atlantic oceanic area
1750–1854
Ships’ logbooks
Daily
Palermo (southern Italy) Greece (northern part) Greece and southern Balkans, partly Cyprus
Documentary sources
Annual
Severe winter extremes Monthly, not continuous
Precipitation (droughts, floods), temperature, extreme
Camuffo & Enzi (1992, 1994a 1995b); Brazdil et al. (1999); Glaser et al. (1999); Pfister et al. (1999) High pressure, low Camuffo & Enzi dispersion (based (1994, 1995c) on dry fogs) Piervitali & Colacino Precipitation (2001) (drought) Precipitation Diodato (2006)
Temperature
Repapis et al. (1989)
Temperature and Xoplaki et al. (2001) precipitation, inclusive droughts, floods Flood levels of the Fraedrich & Bantzer Nile river (1991); Eltahir & Wang (1999); Kondrashov et al. (2005 & references therein) www.ucm.es/info/cliwoc Pressure fields, storm frequencies, wind force, wind direction and general weather
Mediterranean Climate Variability over the Last Centuries: A Review
Italy
37
38 Mediterranean Climate Variability
is increasingly used for regional-to-continental climate reconstructions and analyses of extremes during the last few centuries before instrumental data became available (Pfister, 1992, 2005; Pfister et al., 1998, 1999; Bra´zdil et al., 1999, 2003, 2004, 2005; Glaser et al., 1999; Pfister and Bra´zdil, 1999; Ra´cz, 1999; Bra´zdil and Dobrovolny, 2000, 2001; Glaser, 2001; van Engelen et al., 2001; Benito et al., 2003a,b; Shabalova and van Engelen, 2003; Bartholy et al., 2004; Chuine et al., 2004; Luterbacher et al., 2004; Casty et al., 2005a; Guiot et al., 2005; Glaser and Stangl, 2005; Menzel, 2005; Przybylak et al., 2005; Xoplaki et al., 2005). Documentary evidence is best suited to analyse the impact of natural disasters (e.g. severe floods, droughts, windstorms, frosts, hailstorms, heat waves) on past societies (see Pfister, 1992, 1999, 2005; Martin-Vide and Barriendos, 1995; Barriendos, 1997, 2005; Barriendos and Martin-Vide, 1998; Pfister et al., 1998, 1999, 2002; Bra´zdil et al., 1999, 2003, 2004, 2005; Pfister and Bra´zdil, 1999 for a review and references therein; Barriendos and Llasat, 2003; Benito et al., 2003a,b). Analysing reports of climate extremes in the context of other proxy climate information enables an investigation of the relationship between fluctuations in mean climate and the frequency of extremes (Katz and Brown, 1992) – a major source of societal concern in light of global warming (e.g. Pfister, 2005). Documentary evidence comprises all non-instrumental man-made data on past weather and climate as well as instrumental observations, prior to the set-up of continuous meteorological networks. Non-instrumental evidence is subdivided into descriptive documentary data (including weather observations, e.g. reports from chronicles, daily weather reports, travel diaries, ship logbooks, etc.) and documentary proxy data (more indirect evidence that reflects weather events or climatic conditions such as the beginning of agricultural activities, the time of freezing and opening up of waterways, religious ceremonies in favour of ending meteorological stress, etc. Bra´zdil et al., 2005 and references therein). In general, descriptive evidence has a good dating control and high temporal resolution (often down to the single day). The data distinguish meteorological elements and cover all months and seasons. However, descriptive evidence is discontinuous and biased by the perception of the observer (Glaser, 2001; Bra´zdil et al., 2005; Pfister, 2005). The methods of analysis involves collocating a substantial amount of quality-controlled descriptive and proxy evidence for a given region (Pfister et al., 1998, 2002; Bartholy et al., 2004; Bra´zdil et al., 2005; Glaser and Stangl, 2005; Pfister, 2005; Przybylak et al., 2005). Long series of documentary proxy data are calibrated against instrumental measurements. The spatial and logical comparisons and crosschecking of the entire body of evidence collocated for a given month or season allows the assessment of a climatic tendency, which is in the form of an intensity index for temperature and/or precipitation covering the last centuries. Very recently, Pfister (2005) nicely discussed the climate
Mediterranean Climate Variability over the Last Centuries: A Review
39
sensitivity of early modern economies, climate impacts and crises during the LIA at European scale pointing to the importance of temporally high resolved information from documentary proxy evidence. Figure 3 presents an example of a document that describes the damages experienced in irrigation network of the city Lleida (Catalonia, Spain) on 10 June 1379. Usually these hydraulic installations experienced slight to moderate damages when the snowmelt in the Pyrenees Mountains produced an increase of water flow in Segre River. In this case, the damage was particularly large. Then, not only climatic hazards information can be analysed, but also attitude of human communities in previous historical contexts: People and authorities knew about extreme weather events and accepted a certain risk of damage or destruction. They recorded the phenomena by evaluating the damages and by preparing reconstructions.
Figure 3: Left: An example of a document that describes the damages in irrigation network of the city Lleida (Catalonia, Spain) on 10 June 1379. Right: Painting of a historical flood event in Barcelona, Spain 1862 (Amades, 1984). In the following subsection, we review the availability of documentary evidence from the Mediterranean area and how are they used for local-to-regional climate reconstructions. The compiled information is summarized in Table 1.
Iberian Peninsula Recent studies performed in the last decade have confirmed that the two Iberian countries (Spain and Portugal) have a considerable amount of climate documentary information since the Low Middle Age (fourteenth–fifteenth countries).
40 Mediterranean Climate Variability
Spain, in particular, possesses information with a good degree of continuity and homogeneity for a large number of cities. Thus, the Spanish historical archives exhibit great potential for inferences into climate variability at different timescales and for different territories. Garcı´ a-Herrera et al. (2003a) report on the main archives and discuss the techniques and strategies to obtain climaterelevant information from documentary records. Precipitation patterns are the most evident limiting factor for human activities and natural ecosystems in the Mediterranean. If climatic change produces quantitative or qualitative alteration of rainfall patterns, human communities and natural ecosystems can be irreversibly damaged. Martin-Vide and Barriendos (1995), Barriendos (1997, 2005) and Barriendos and Llasat (2003) used rogation ceremony records from Catalonia (Spain) for precipitation reconstructions (Fig. 4). Rogations were an institutional mechanism to drive social stress in front of climatic anomalies or meteorological extremes (e.g. Barriendos, 2005). Municipal and ecclesiastical authorities involved in the process guarantee the reliability of the ceremony and continuous documentary record of all rogations convoked. On the other hand, duration and severity of natural phenomena-stressing society is perceived by different levels of
Figure 4: Left: Manuscript recording one severe rogation ceremony following drought in Girona city (Catalonia, NE Spain), April 1526. Historical City Archive of Girona. This ceremony consists in a pilgrimage from Girona to St. Feliu de Guixols (20 km) to immerse one relic in sea water (mummified head of Saint Feliu). Right: Saint Narcisus was the first bishop of Girona. His relics were displayed in the Cathedral in rogations both for drought (pro pluvia) and during long periods of rain (pro serenitate). (Martı´ n-Vide and Barriendos, 1995; Barriendos, 1997).
Mediterranean Climate Variability over the Last Centuries: A Review
41
liturgical ceremonies applied (e.g. Martı´ n-Vide and Barriendos, 1995; Barriendos, 1997, 2005). Rodrigo et al. (1998) analysed climatic information in private correspondence of a Jesuit Order in Castille (Spain) for 1634–1648. They showed prevalence of intense rainfall and cold waves in that period. Rodrigo et al. (1999, 2000, 2001) reconstructed a 500-year seasonal precipitation record for Andalusia (Spain) and derived a winter North Atlantic Oscillation (NAO) index based on meteorological information on droughts, abundant rainfall, floods, hail, etc. This information was obtained from a wide variety of documentary sources such as Municipal Acts, private correspondence, urban annals, chronicles, brief relations describing extreme events, agricultural records, etc. Results of these studies indicate rainfall fluctuations, without abrupt changes, in the following alternating dry and wet phases: 1501–1589 dry, 1590–1649 wet, 1650–1775 dry, 1776–1937 wet and 1938–1997 dry. Possible causal mechanisms for these variations most likely include the NAO with drought (floods) being related to extreme positive (negative) NAO values. Precipitation in the Canary Islands has been reconstructed using agricultural records for the period 1595–1836 (Garcı´ a-Herrera et al., 2003b). Barriendos and Martin-Vide (1998), Benito et al. (2003a,b) and Llasat et al. (2003) investigated flood magnitude and frequency within the context of climatic variability for the last centuries for central Spain and Catalonia. The authors found evidence for high flood frequencies in the past, which are similar to present conditions. Comparable catastrophic events have been recorded at least once each century. Most recently, weather information was obtained from original documentary sources from the northeastern (Barcelona) and southeastern (Murcia) coast of the Iberian Peninsula, respectively (Barriendos and Rodrigo, 2005). The climatic indicators used are ‘‘pro pluvia’’ (ceremonies to obtain rainfall) and ‘‘pro serenitate’’ (ceremonies to stop continuous rainfall events) rogations. These proxy data records offer highest reliability and excellent perspectives in historical climatology for the Roman Catholic cultural world (Martı´ n-Vide and Barriendos, 1995; Barriendos, 2005). A numerical index, ranging from 3 (severe droughts) to þ 3 (catastrophic floods), was established to characterize the seasonal rainfall and its evolution. The information was calibrated against overlapping instrumental data (for Barcelona 1786–1850; in case of Murcia 1866–1900), whereas a cross-validation procedure was employed to confirm the reliability of the calibrations. The regression equations between index values and instrumental data were used to extend seasonal rainfall series for the Iberian Peninsula back in time (Rodrigo et al., 1999; Barriendos and Rodrigo, 2005). Figure 5 presents standardized anomalies of seasonal rainfall in Barcelona and Murcia with regard to the 1961–1990 reference period. It shows the fluctuating character of precipitation with important wet (first half of the
Rainfall anomalies
Winter (Barcelona)
Rainfall anomalies
Spring (Barcelona)
Rainfall anomalies
Autumn (Murcia)
Figure 5: Top: Reconstruction of winter precipitation anomalies at Barcelona, Middle: Spring at Barcelona and Bottom: Autumn at Murcia back to the sixteenth century. Solid thin line: Seasonal rainfall standardized anomalies (reference period 1961–1990); Thick line: 11-year running average (Barriendos and Rodrigo, 2005).
Mediterranean Climate Variability over the Last Centuries: A Review
43
seventeenth century and around 1850) and dry periods (around 1650 and 1750) during winter (Barcelona). For spring, a wet period in the last decades of the sixteenth century and a dry period from the first decades of the seventeenth century to approximately 1750 has been found for Barcelona. In Murcia, there is a slight decreasing trend, visible in autumn rainfall. The earliest documentary sources for Portugal (Western Iberia) correspond to the seventeenth century. Daveau (1997) analysed private letters of a priest of the Jesuit Order (Anto´nio Vieira) and has reconstructed weather in Central Portugal from December 1663 to September 1665. It is stated that between December 1664 until March 1665 very long sequences of rainy weather occurred, as well as floods of the large Iberian (Tagus) and Portuguese rivers (Mondego). Concerning temperature variability, for example, Vieira writes that in April 1664 there occurred ‘‘the greatest cold as in December (in the first half of the month) as well as hot spells similar to those in Guinea (western Africa) (at the end of the month)’’. Alcoforado et al. (2000) have used several documentary sources such as diaries, ecclesiastical documents (including references to ‘‘pro pluvia’’ and ‘‘pro serenitate’’ rogation ceremonies, see above) to reconstruct temperature and precipitation variability in southern Portugal, during the Late Maunder Minimum (LMM, 1675–1715). One of the diaries refers to the period from 1696 to 1716 and although non-meteorological news is the thread throughout it, there are detailed descriptions of weather and of the author’s perception of its consequences (Alcoforado et al., 2000). The main conclusions are that, after 1693, conditions in Portugal were rather cold (with snowfall events in Lisbon that hardly ever occur nowadays). Precipitation, on the other hand, showed a very pronounced variability, similar to the present. Taborda et al. (2004) extended the study covering the whole eighteenth century based on descriptive documentary sources (institutional, ecclesiastical, municipal, private and from the press) and early instrumental records (from 1781 to 1793). The winters of 1708–09 (also described in Alcoforado et al., 2000), 1739–40 and 1788–89 were particularly severe. All these winters coincide with very cold conditions on the European scale (Luterbacher et al., 2004). In the beginning of the eighteenth century, precipitation shows strong variability (Fig. 6) with persistent rain from winter 1706–07 until summer 1709 and droughts during winter 1711–12 and between spring 1714 and autumn 1715. Very strong rainfall variability characterized the 1730s confirmed by the highest frequency of ‘‘pro pluvia’’ and ‘‘pro serenitate’’ rogation ceremonies. At the end of the eighteenth century, a period of eight rainy years, beginning in 1783 has to be mentioned. A significant, although low negative correlation, was found between the North Atlantic Oscillation Index (NAOI; Luterbacher et al., 1999, 2002a) and yearly and seasonal precipitation in Portugal (Fig. 6).
44 Mediterranean Climate Variability
Figure 6: Annual precipitation index in southern Portugal (Taborda et al., 2004).
Regular meteorological observations in Lisbon begun in December 1815, carried out by M.M. Franzini, due to public health needs (Alcoforado et al., 1997, 1999). Although 1815–1817 meteorological data were published by the Portuguese Academy of Sciences, the subsequent data have been gathered mostly from newspapers. The data between 1817 and 1854 presents some gaps, one of them lasting nine years (1826–1835). The data were used to study the relation between climate and society (agriculture, human health, necrology). An attempt to construct a reliable meteorological series for Lisbon from 1815 to the present is on its way (M. Alcoforado, personal communication). In summary, climatic research from documentary sources in the Iberian Peninsula is still in its early stages. The current knowledge provides a patchy vision of past climate in Spain and Portugal. Mostly because of lack of funding, there has not been a systematic attempt to explore the main Spanish and Portuguese archives, such as the Archivo de Simancas, and a lot of local archives. Despite the effort required to collect information from original archives, a tremendous unexploited potential in different archives in many Spanish regions exist: economic information from agriculture (production and tributary statistics), monastic documentary sources for High Middle Age (eighth– thirteenth centuries), or documentary testimonies from the top of ecological thresholds (farming on medium/high mountains). Most of the information is related to the rainfall, which is of greatest importance for an agricultural economy. In this sense, not much research on temperature has been made so far. The Spanish groups working in the paleoclimatology field have a network called RECLIDO (Climate Reconstruction from Documentary sources; www.ucm.es/ info/reclido), which summarizes most of the work done in Iberia. In Portugal, research is being carried out to collect data from the sixteenth and seventeenth
Mediterranean Climate Variability over the Last Centuries: A Review
45
centuries, mostly referring to rainfall and its consequences on agriculture (J. Taborda, personal communication).
France Pichard (1999) has studied the variations of climate and hydrology in southern France from documentary sources. They are based on records of extreme events like floods, presence of ice, insect invasion, long instrumental records and economic data. Figure 7 shows as an example the statistics of floods in the Durance Valley, a river running from the Alps to the Rhone Valley in Avignon. This river was well known in the previous centuries for frequent flooding events. It appears that the period 1540–1900 was characterized by much more frequent floods as compared to the twentieth century. The same kind of situation is also reported for the Rhone Valley. Floodings in the Durance catchment reflect mainly winter and spring precipitation, the most dominant in the southern Alps. Before 1540, only the period 1330–1410 was wet. It is assumed that these two wet periods were due to higher winter precipitation and also much more snow in the southern Alps.
Years ( AD)
Figure 7: Reconstruction of the Durance river (France) flood events by Guilbert (1994) reconstructed from documentary sources: number of months with floods per decades. As for other parts of the Mediterranean (Xoplaki et al., 2001 for Greece), the wettest climate of the ‘‘Little Ice Age’’ (LIA) occurred during the periods 1650–1710 and 1750–1820. Guiot et al. (2005) recently reconstructed the temperature at Marseille Observatoire (Fig. 8) based on a combination of a variety of documentary proxy evidence (among them grape-harvest dates from France) and tree-ring information. They showed a LIA cooling of 0.5 C with maxima of 1.5 C, in phase with western Europe. The recent warming of
46 Mediterranean Climate Variability
Figure 8: Reconstruction of the summer (April–September) temperature in Marseille Observatoire compared with observations (all the values are expressed in C as departures from the 1961–1990 mean of 19.5 C). The reconstructions are obtained from tree-ring series combined with documentary data (vine-harvest dates, number of months with ice in the Rhone River etc.). In grey is the confidence interval at the 90% level (Guiot et al., 2005, the last millennium summer temperature variations in western Europe based on proxy data, Holocene 15,489–500). Copyright 2005, Edward Arnold (Publishers) Limited.
1 C was never reached in the context of the last millennium even if it still lies within the confidence interval of the previous centuries (Guiot et al., 2005). The Swiss physicist Louis Dufour (1870) was the first to discover the value of dates on the opening of vine harvests for the reconstruction of temperatures in the pre-instrumental period. He was followed by the French climatologist Alfred Angot, who provided a catalogue of documentary evidence in France (Angot, 1885). French records of grape-harvest dates in Burgundy (Le Roy Ladurie and Baulant, 1980, 1981; Pfister, 1992; Soriau and Yiou, 2001; Chuine et al., 2004; Le Roy Ladurie, 2004, 2005; Menzel, 2005) were used to reconstruct spring–summer temperatures from 1370 to 2003 using a process-based phenology model developed for the Pinot Noir grape (Chuine et al., 2004). The results reveal that summer temperatures as high as those reached in the 1990s have occurred several times in Burgundy since 1370. However, the summer of 2003 appears to have been extraordinary, with temperatures that were probably higher than in any other year since 1370 (Chuine et al., 2004). Le Roy Ladurie (2004) recently presented a very nice overview of the past climate conditions, socio-economic conditions and climate impacts in France. Summarized historical written documents in France are insufficiently exploited. After the work of Le Roy Ladurie (1983), grape-harvest dates series have shown their potential and recent work of Chuine et al. (2004) and Le Roy Ladurie (2005) has proved that it is possible to translate them into quantitative temperature series. However, this has been limited to the non-Mediterranean part of France and Switzerland, even if potentialities for an extension exist
Mediterranean Climate Variability over the Last Centuries: A Review
47
(Pichard, 1999). This latter author has shown that other sources are also available, such as religious processions for rain (such as from the Iberian Peninsula, see above), cereal price series, insect invasions, river floodings and ice presence. The French corpus of documentary sources opens new horizons for climate historians due to a long tradition of centralizing policy in the French Kingdom. Thus, it exhibits great potential for inferences into climate at different timescales and for different regions. The ‘‘Intendances’’, for example, were areas for ‘‘Police, Law and Finance’’ in the seventeenth and eighteenth centuries. Their archives (Garnier, 2005) contain a lot of grievances, tax exemption for floods, hailstorms, dryness and storms. Likewise, the sources of ‘‘Admiralties’’ (‘‘Amiraute´s’’ in French), naval areas, created in 1665 for French Mediterranean coastline are very interesting for their ‘‘Ship Log reports’’ (Garnier, 2004). These latest papers are very valuable because of their weather and chronology observations (wind directions, storms, shipwrecks, dates). Religious records must not be forgotten either with numerous ex voto of storms or floods, rogations and monastic documentary sources (account books). Finally, it is necessary to take account of the vast amount of agricultural records like diaries of page`s (fifteenth and nineteenth centuries) – fluent peasants in Catalonia and Languedoc – and the books of Parceria – farming leases studied by Le Roy Ladurie (1966) and Garnier (2005b). All these proxy sources have their own limitations and biases (for example productivity improvements in agriculture, variations of the river depth due to sediment transfer, etc.). An integrated approach involving many proxies (Guiot et al., 2005) or in combination with modeling as used by Chuine et al. (2004) for grape-harvest dates might be a successful way for further research in the area. In the future, the French research in climate history should enjoy a new boom with the national programme OPHELIE (Observation PHEnologiques pour reconstruire le CLImat de l’Europe) that brings together mathematicians (P. Yiou) and historians (E. Le Roy Ladurie and E. Garnier). This ambitious project has been put in charge of the Laboratory of Climatic and Environmental Sciences CEA CNRS of Gif-sur-Yvette and it aims at proposing a climate reconstruction based on phenological data as grape-harvest dates, on municipal account books, rogations and an international meteorological network established by Vicq d’Azyr at the end of the eighteenth century.
Italy Italy has a long history with early civilization and written documents beginning in the Roman times. For instance, it was possible to reconstruct the flooding series of the river Tiber for Rome back to 2400-year BP (Camuffo and Enzi, 1995b, 1996). The flooding of river Tiber in Rome offers the opportunity of having one of the longest discharge series in the world. Over the centuries, the river response has had some minor changes to the meteorological forcing
48 Mediterranean Climate Variability
derived from the changes affecting the territory, the banks and urban development in Rome. However, the most important change was the rise of the banks in 1870, which practically reduced, or even stopped the series of floods. Floodings mainly occur in winter (especially November–December). Strong precipitation events reflect both the air–sea temperature contrast and the occurrence of the Scirocco wind. The Tiber had two major periods of increased overflowing frequency at the beginning of the Spo¨rer Minimum and at the onset of the Maunder Minimum, i.e. 1460–1500 and 1600–1660. The periods 1400–1460, 1500–1600 and 1660 onwards show a very low rate of flooding, which was further reduced after the works in 1870. Roman literature reports also on major or impressive events that happened more or less in all regions of Italy (and some in Europe too). Part of this information is related to wars or other political or social events, which may affect the objective description. The abundance of the data decreases in the Medieval period, when the social conditions were very bad. Starting with 1000, the improvement in the social conditions were reflected in a second flourishing of the culture, which culminated in 1400–1500. However, during the 1100–1200 period, a number of cities started to fight against the Emperor and other authorities (especially in the north of Italy). With independence, people started to document extreme events, natural hazards, yields, etc. described in annals and chronicles. In 1300, Florence had free schools and all citizens were able to write and read. The 1400–1500 period, flourishing in 1500 with Leonardo da Vinci, Galileo and others, is rich in literary and historical culture and is the background for science as well. Thus, Italian archives, libraries and museums provide a great number of written historical sources on different aspects of past climate reaching back more than 1000 years (e.g. Camuffo, 1987). Over the centuries, many subjective reports on extremely cold winters can be found. Fortunately, this abundant information can objectively be evaluated per classes of severity. A ‘‘great winter’’ was defined when the cold was particularly severe over a large area causing well-documented exceptional events, e.g. large water bodies were frozen, with ice sustaining people and chariots, wine was frozen in butts, death of people, trees and animals. The term ‘‘severe winter’’ was used when people, trees and animals were killed, and only minor rivers were frozen over. ‘‘Mild winters’’ when ice was missing and plants had early growing and flowering. In northern Italy, the freezing of the Venice Lagoon and its deeper canals was a very useful reference to quantitatively establish the degree of severity over the centuries. This information was based on a large number of citations, pictorial and literary representations (Camuffo, 1987, 1993; Camuffo and Enzi, 1992, 1994a,b, 1995a,c; Enzi and Camuffo, 1995a; Camuffo and Sturaro, 2003; Fig. 9). Freezing was particularly frequent in the 1400–1600 period and then 1700–1850. The coldest winter in the series was 1708–09 (most probably
Mediterranean Climate Variability over the Last Centuries: A Review
49
Figure 9: 1,000 year chronology of the Venice Lagoon freezing (Camuffo, 1997). the coldest winter in Europe for at least half a millennium, Luterbacher et al., 2004). Other very cold winters were experienced in 1928–29 and 1788–89 (Camuffo, 1987). Past flooding in Venice is another important factor to understand regional climate variability. These ‘‘High Water’’ (Camuffo, 1993; Enzi and Camuffo, 1995; Camuffo et al., 2005) occur when the sea level rises more than 110 cm above the mean level (with respect to the yearly average level observed in 1897). They caused enormous problems to the city. For this reason, High Waters were reported in public and private documents (regular instrumental records, i.e. tide gauge, began in 1872). The problem nowadays is dramatically relevant because of the damage to historical buildings and monuments. Flooding surges are due to a cyclonic circulation moving over western Mediterranean: the Scirocco wind is strong and drags water to Venice; the corresponding pattern of atmospheric pressure over the Mediterranean further displaces water towards Venice. The sea level rise is increased or decreased by further factors such as luni-solar forces, free oscillations (seiches) in the Adriatic basin, and an additional sea level rise due to global warming. In the past, deep cyclonic circulations generating High Waters in Venice were particularly frequent in the first decades of 1500 and at the turn of the eighteenth century. Nowadays, the surge frequency has increased exponentially due to the combined effect of soil subsidence and sea level rise. The relative sea level change in Venice is a vital problem for the city, which has raised 61 11 cm over the last few centuries.
50 Mediterranean Climate Variability
The brown belt of the algae which live in the tidal range and the upper front is a good indicator for the average high tide level. This indicator was accurately drawn by Antonio Canaletto (1697–1768) and Bernardo Bellotto (1722–1780) in their ‘‘photographic’’ paintings (Camuffo et al., 2005). In northern Italy, the long series of locust invasions constitutes an index of the frequency of easterly circulation in the summertime, which transported swarms originated in the Pannonian plain (Hungary) (Camuffo and Enzi, 1991). Locusts from Anatolia or the Near East infested the Pannonian plain. In the summertime, eastern winds of Bora type, transported the swarms westwards, i.e. from Hungary to northern Italy. Annals and chronicles report the list of the damaged areas, often followed by famine and epidemics, and it is often possible to follow the path and spread of swarms. Locust invasions (and cold summer inflows) were more frequent in the mid-fourteenth century, during the Spo¨rer Minimum (1460–1500), with a major peak in the early sixteenth century, and at the very beginning of the Maunder Minimum (1645). Invasions were terminated when the Pannonian plain was densely cultivated and the locust eggs were destroyed. In Sicily and the western coast of Italy, locust invasions were mainly related to southerly winds (mainly Scirocco) that transported swarms from northern Africa, e.g. in 1566/1572. In southern Italy, parts of the semi-arid territory was left uncultivated and used for grazing sheep, so that it was naturally infested with locusts. The severity of the plague was determined not only due to climatic factors, but also by the effectiveness of the methods used to fight them. Intensive land cultivation was the most effective system that terminated this plague in the first part of the nineteenth century. Piervitali and Colacino (2001) analysed drought events that occurred in western Sicily during the period 1565–1915 using information on religious processions performed in the small town of Erice (Italy). Together with other documents and manuscripts found in the city library, a drought chronology was reconstructed. Diodato (2006) used a variety of different written records of weather conditions that affect agriculture and living conditions as a proxy for instrumental annual precipitation for the Palermo (southern Italy) region 1580–2004. Data were compiled from various kinds of documents, including chronicles, annals, archival records and instrumental data (from 1807 onwards). Diodato (2006) found a wet period between 1677–1700 that seem to be in agreement with findings from Greece (Xoplaki et al., 2001) and Spain (Barriendos, 1997). A wetter phase has been found between 1769 and 1922 followed by a general decrease until the end of the twentieth century. Intense natural pollution in Italy occurred in the past due to the intense volcanic activity, which has diminished in recent times. Between 1500 and 1900, the Mediterranean volcanoes Etna, Vesuvius, Vulcano and Stromboli were
Mediterranean Climate Variability over the Last Centuries: A Review
51
particularly active and caused the so-called dry fogs (Camuffo and Enzi, 1994, 1995a). Acid volcanic fogs consist of a more or less dense mist composed of gasses and aerosols with reddish colour and foul smelling. This mist is dry. The most dramatic episode occurred in 1783, due to Icelandic volcanic activity at Laki, which affected most of the NH (Franklin, 1784). In Italy, this phenomenon appeared most frequently from spring to early summer when the volcanic emissions were less easily dispersed in the atmosphere. Two main factors are prominent: the Mediterranean Sea was relatively cold giving rise to very stable atmospheric conditions and low dispersion potential. Further, the Azores anticyclone extended over the Mediterranean, which reduced winds. Under such conditions the volcanic emissions, especially those emitted at low levels, remained entrapped in the stable boundary layer, which were then transported towards the land by a gentle breeze. The dry fogs persisted for days or weeks. From the analysis of these pollution episodes, which have occurred in the Po Valley over the last millennium, it is difficult to identify the individual volcano that has determined the occurrence of each dry fog event. Volcanic clouds crossed Italy from south to north, destroying from one-third to one-half of the maize or wheat yield. It would seem much more reasonable to note that the phenomenon became frequent only after Stromboli became active once again. In agricultural meteorology of the 1800s, the phenomenon was so relevant that sources distinguished between the caustic dry fogs that damaged the vegetation and damp fogs, with positive effect because they act as a nutrient. From 1300 to 1900, some 50 anomalous fog events have been noted, 30 of those were certainly corrosive, i.e. of volcanic nature. The frequency of these events culminated between the middle of the 1700s and the middle of the 1800s. Summarized, the documentation of past climate and extremes in Italy based on documentary evidence is widespread and reliable. The South had different political vicissitudes, but in any case it had a very flourishing culture (e.g. Sicily, Apulia, Naples), with a similar amount of climate and weather descriptions. A number of people wrote diaries, logs, reports etc. so that the documentation becomes abundant and sometimes even quantitative. There is still much potential in Italy to collect, read and digitize these climate-related information as the State Archives of Italy, for instance have shelves extending 12,000 km, the State Archive in Venice has 17 km shelves (public and private libraries, monasteries, private collections etc. handwritten and printed documents of any type not included).
Southern Balkans, Greece and Eastern Mediterranean The Balkan Peninsula (Greece, former Yugoslavian countries, Albania, Bulgaria and Romania) provides rich archives of documentary data. Repapis et al. (1989) investigated the frequency of occurrence of severe Greek winters based on
52 Mediterranean Climate Variability
evidence from monastery and historical records during the 1200–1900 period. They found evidence, that the coldest periods occurred in the first half of the fifteenth century, in the second half of the seventeenth century and in the nineteenth century. Grove and Conterio (1994, 1995), Grove (2001) and Xoplaki et al. (2001) reported on the variability of climate and extremes (severe winters, droughts and wet periods) during parts of LIA and its impact on human life using different kind of written sources. Figure 10 presents the estimated winter temperature and precipitation conditions for Greece during the LMM 1675–1715 period based on documentary proxy evidence (Xoplaki et al., 2001). Xoplaki et al. (2001) found, that during these periods more extreme conditions were apparent compared to the late twentieth century. Xoplaki et al. (2001) extended the analysis for the 1780–1830 period. Documentary information on ‘‘Medieval Warm Period (MWP)’’ and the beginning of LIA in eastern Mediterranean is provided by Telelis (2000, 2004).
Figure 10: Top: Averaged winter temperature and Bottom: Precipitation conditions for Greece estimated from documentary proxy evidence during the LMM (Xoplaki et al., 2001). Values of plus (minus) 3 indicate exceptionally warm (cold) in case of temperature and wet (dry) for precipitation. An index of 0 for a particular winter means normal conditions compared with the 1901–1960 average.
Mediterranean Climate Variability over the Last Centuries: A Review
53
Compared to the wealth of data found in central and western Europe the data density for the southern Balkans, Greece and eastern Mediterranean for the last few centuries is rather low. This may be attributed to the Turkish occupation, which lasted from the fifteenth to the nineteenth century (Xoplaki et al., 2001). However, it is believed that a number of important monastery memoirs covering the last at least half a millennium show evidence of past weather and climate and optical phenomena from important tropical volcanic eruptions (C. Zerefos, personal communication). It is assumed, that there are detailed documentary data available from the Turkish archives that could be explored and used for climatic reconstructions. Further, it is also believed that early instrumental data from Cyprus, Syria and Greece starting in the eighteenth century and from Egypt and Malta from the early nineteenth century can be obtained from different sources.
Northern Africa There is only very limited climate information available so far from northern African countries based on documentary evidence. A notable exception is the record of the flood levels of the Nile river, which was analysed by hydrologists, climatologists and historians. There are a number of studies dealing with the reconstruction and analysis of the data from written records (Fraedrich and Bantzer, 1991 and references therein; De Putter et al., 1998; Eltahir and Wang, 1999; Kondrashov et al., 2005). Pharaonic and medieval Egypt depended solely on winter agriculture and hence on the summer floods. The rise of the waters of the Nile was measured therefore regularly from the earliest times (e.g. Eltahir and Wang, 1999; Kondrashov et al., 2005 and references therein). Several authors compiled the annual maxima and minima of the water level recorded at nilometers (generally an instrument that measures the height of the Nile waters during its periodical flood) in the Cairo area, in particular at Rodah Island, from 622 to 1922. There is evidence of low Nile floods occurring in the periods 1470–1500, 1640–1720 and a number of low floods from 1774 to 1792 (Fraedrich and Bantzer, 1991). Ship Logbooks as a New Documentary Proxy for Past Mediterranean Climate Weather observations have been made on-board sailing ships as part of a daily routine since the mid-seventeenth century. Procedures for marine observations were not, however, formalized until the International Maritime Conference of 1853 (Maury, 1854). The lack of consistency of the records before this date might help to account for the reluctance of climatologists to exploit the earlier records in any comprehensive fashion. Recent studies of ships’ logbooks for the period 1750–1850 undertaken as part of the CLIWOC project (Climatological Database for the World’s Oceans, CLIWOC Team, 2003; Gallego et al., 2005;
54 Mediterranean Climate Variability
Garcı´ a-Herrera et al., 2005a,b; Jones and Salmon, 2005; Wheeler, 2005; www.ucm.es/info/cliwoc) and for the period 1680–1700 (Wheeler and Sua´rez Domı´ nguez, 2006) have, however, demonstrated the value of such material as a source of reliable climatic data and information. Studies have also confirmed the availability of a large number of such logbooks for the Mediterranean. After 1850 most ships provided instrumental data, but such provision is exceptional for the years before the mid-nineteenth century. The climatic information contained in these early logbooks falls under three headings: those of wind force, wind direction and general accounts of the weather. The layout of logbooks varied slightly within and between nations but they all contain much the same information, and the presentation exemplified in Fig. 11 is typical. Wind direction and force were recorded at noon each day, these observations often being supplemented by additional records at other hours providing an unrivalled picture of short-term variation. Observations were also included on such things as the state of the sea, cloudiness, visibility and the incidence of particular phenomena such as rain, snow, thunder and fog. Although based on visual observations and individual judgment, these estimates were made by experienced officers whose abilities would differ little from those of today’s deck
Figure 11: An example of a Spanish logbook page from 1797 (Courtesy of the Archivo del Museo Naval (Archive of the Naval Museum), Madrid, Spain).
Mediterranean Climate Variability over the Last Centuries: A Review
55
officers, many of whom continue to make similar records in the logbooks of merchant and military vessels many of which are used by the forecasting services. Each of the early records is presented in narrative, non-numerical form. In that sense they differ from the instrumental data gathered in such sources as ICOADS (International Comprehensive Ocean and Atmospheric Data Set, Worley et al., 2005) although they do occasionally provide temperature and barometric data, some from as early as the late eighteenth century. These narrative data, written in the language and vocabulary of the age (and nation), need to be transformed into present-day terms (and English) before they can be subjected to scientific analysis. The CLIWOC project has established procedures and methods whereby these transformations can be made. The project has also assessed the intrinsic reliability of the original observations (Wheeler, 2005). These activities have permitted the construction of a database (Ko¨nnen and Koek, 2005) containing quality-controlled data for the equivalent of 280,000 days of observations. Figure 12 shows the geographic range and coverage of these CLIWOC data. To date, logbook-based studies of the Mediterranean climate have been limited to the geographically peripheral area of the Strait of Gibraltar, and to particular historical events (Wheeler, 1987, 2001). Nevertheless such undertakings have amply demonstrated the advantages of using these data to reproduce daily and seasonal synoptic patterns. These exercises have also demonstrated that such sea-based data can be profitably articulated with those from land stations and do not stand apart as a data set. The CLIWOC project was focused on major oceanic regions and excluded all enclosed seas such as the Mediterranean.
Figure 12: CLIWOC version 1.5 data coverage. Each dot represents a daily observation.
56 Mediterranean Climate Variability
There is, however, no shortage of logbooks for the region. Remarkably, the majority of these are found not in the archives of Mediterranean states but in the United Kingdom. British political strategy has been based on sea power from the seventeenth century and as long ago as 1680 British warships and fleets were active in the area. With the establishment of bases, particularly in Gibraltar and, though more temporarily, Port Mahon, British interest in the western Mediterranean was to persist, unbroken, for three centuries. It has been estimated (D. Wheeler, personal communication) that for the Mediterranean there are the equivalent of over 1,000,000 days of data to be extracted from British logbooks of which 4,000–5,000 are estimated to exist in UK archives (principally in The National Archives in Kew, South West London) over the period from 1680 to 1850. From 1700 onwards the record is probably unbroken, with at least one fleet or squadron active somewhere in the Mediterranean at any given time. The number of logbooks varies according to the political climate (war time yields far more records than periods of peace) and Fig. 13 summarizes their decadal availability. The geographic range of currently available observations is by no means restricted to the British and allied ports. Vessels were based in Naples, Cyprus and Alexandria at different times, and British naval policy required Royal Navy ships to cruise extensively providing thereby something close to the observational network offered by today’s merchant services. Given that military action was frequently necessary against the North African Barbarian States,
Figure 13: Decadal distribution of logbooks in British archives from the Mediterranean region for the 1670–1840 period. The graph shows the number of ships and logbooks estimated from 1670 to 1850.
Mediterranean Climate Variability over the Last Centuries: A Review
57
there exists also the opportunity partly to fill the gap noted on a number of occasions in this chapter that prevails over this most southerly sector of the region. It is not known if further archival sources in such historic centres as Venice, Istanbul or Alexandria might yield additional logbooks or similar documents, although S. Enzi and D. Camuffo have confirmed the existence of some logbooks in Italian archives. Further logbook collections are also known to exist in French and Spanish Archives. The French logbooks cover the period 1671–1850. Most of them were prepared during coastal voyages to Spanish or Italian ports (Garnier, 2004). It is estimated that approximately 500 of such logbooks are preserved but have remained undigitized. A further 100 logbooks are preserved in Spain, corresponding mostly to coastal sailing by ships of different Catalonian companies (R. Garcı´ a-Herrera personal communication; Prohom and Barriendos, 2004). To summarize, logbook data offer a number of significant benefits to the climatologists: Firstly, the data are fixed by time and location, being recorded each day at mid-day, with a further note that includes the ship’s latitude and longitude. Secondly, the observations are homogenous in that they are recorded using a widely adopted vocabulary and based on a set of common practices that prevailed even during those years and decades before adoption of the Beaufort system. Thirdly, the data should not be regarded as ‘‘proxy’’: they constitute first-hand and direct observations on the weather at the time. Fourthly, and very importantly, they are the only such source of information for the oceanic and sea areas. This is of significance in the Mediterranean as the region, where the sea represents a significant proportion of its surface area. Fifthly, logbooks provide information that extends back to the late seventeenth century and express therefore conditions at a critical time of climatic evolution that includes the closing decades of the LMM. And, finally, these data are so abundant, that there is a genuine possibility of providing a daily series from 1700 onwards, especially for the western Mediterranean area.
1.2.2. Evidence from Natural Proxies In this section, we describe natural proxies that are used to reconstruct climate conditions for sub-Mediterranean areas. They include high-resolution proxies such as tree rings, speleothems and corals, but also lower resolution natural proxies such as borehole data. In addition, we discuss new marine archives (non-tropical coral, deep-sea corals) that show much potential for regional climate reconstructions. Table 2 summarizes the climate evidence described in detail in Section 1.2.2. The first part deals with the climate evidence in the different areas of the Mediterranean based on tree-ring data, followed by descriptions of speleothems
Location
Time
Archive
Temporal resolution
Parameter reconstructed
Reference
Northern Spain
1500–present
Tree rings
Seasonal
Saz (2004)
Spain
1000–present
Tree rings
Seasonal
Temperature, precipitation Temperature, precipitation
Galicia (Spain)
1500–present
Tree rings
Annual
Adriatic coast (Italy) Mediterranean Basin Mediterranean Basin Morocco Morocco (regional) Morocco
1653–1985
Tree rings
Winter
1000–2000
Tree rings
Annual
1700–2000
Tree rings
1000–1979 1100–1979
Tree rings Tree rings
1429–2000
Tree rings
April– September Annual October– September Winter
1600–1995
Tree rings
October–May
Southern Jordan
Temperature, precipitation Temperature
Creus-Novau et al. (1997); Saz & Creu-Novau (1999) Creus-Novau et al. (1995) Galli et al. (1994)
Temperature, precipitation Temperature
http://servpal.cerege.fr/ webdbdendro/ Briffa et al. (2001)
Precipitation Precipitation
Munaut (1982) Till & Guiot (1990)
NAO, precipitation Precipitation, drought
Glueck & Stockton (2001) Touchan et al. (1999)
58 Mediterranean Climate Variability
Table 2: Compilation of temporally ‘‘high resolved’’ natural proxies from the Mediterranean (see Section 1.2.2 for details).
Tree rings
Drought
Precipitation Index
Touchan et al. (2005a)
1400–2000
Tree rings
May–August
Precipitation
Touchan et al. (2005b)
300–2001
Proxy record from 18O analysis of a varved lake sediment sequence River sediments
Annual
Summer evaporation (or drought)
Jones (2004); Jones et al. (2004, 2005)
Instantaneous maximum discharge
Flood magnitude/ frequency/extremes
Benito et al. (2003)
1887–1995
Tree rings
Annual
Black Sea (Turkey) Southern central Turkey Sivas (Turkey)
1635–2000
Tree rings
Annual
1689–1994
Tree rings
Annual
1628–1980
Tree rings
Annual
Southwestern Turkey The land area of most Turkey and adjoining region Eastern Mediterranean Central Anatolia Turkey, (Nar Go¨lu¨: 38 270 3000 E; 38 22’30’’N; 1363 masl)
1339–1998 1776–1998 1251–1998
Tree rings
3000 BP –present
Akkemik & Aras (2005) D’Arrigo & Cullen (2001) Touchan et al. (2003)
59
Catalonia (Spain) 40 N–43 N, 0 E–4 E
Akkemik et al. (2005)
Mediterranean Climate Variability over the Last Centuries: A Review
Akkemik (2000)
May–June
Positive correlation between Jan–Feb and Jul–Aug precipitation and tree rings; positive correlation between Mar–Apr temp but negative correlation with May–Jul temp Mar–Jun precipitation Apr–Aug precipitation Feb–Aug precipitation Precipitation
Istanbul (Turkey)
(Continued)
Location
Time
Archive
Temporal resolution
Parameter reconstructed
Reference
Italian Alps
1650–1713 and 1798–1840 2000 BP–present 950–present
Speleothems (growth rate)
Annual
Temperature
Frisia et al. (2003)
Annual
Temperature
Mangini et al. (2005)
Seasonal to annual
Precipitation
Antonioli et al. (2003); Montagna (2004)
Bimonthly
SST, hydrologic balance
Felis et al. (2000, 2004); Rimbu et al. (2001, 2003a)
30–50 years
SST, Sea level
Silenzi et al. (2004)
Sea-water chemistry Biological productivity
Montagna (2004); Montagna et al. (2005a,b)
Central Alps Sardinia (Italy) Northernmost Red Sea
Intertidal area in warmer Mediterranean waters From 200 m to 1200 water depth in the Mediterranean Sea
Speleothems/ 18O Speleothems/ 18O, trace elements 1750–1995 Annually banded 98 year at reef corals, 2.9K year ago (oxygen isotopes, 44 year at Sr/Ca) 122K year ago 1400–present Vermetids (Dendropoma petraeum)
1950–present Deep-Sea corals Seasonal to (Lophelia pertusa, annual Madrepora oculata and Desmophyllum dianthus)
60 Mediterranean Climate Variability
Table 2: Continued.
CentralNorthern Italy
1750–1980s/ 1990s
Evora (Portugal)
1700–1997
Morocco
Last 100 –300 years
8 Boreholessubsurface temperature profiles 1 Boreholesubsurface temperature profile 1 Boreholesubsurface temperature profile
Monthly to weekly
SST Sea-water chemistry
Silenzi et al. (2005)
Secular temperature trends
Ground surface temperature variations (after inversion) Ground surface temperature variations (after inversion)
Rajver et al. (1998)
Ground surface temperature variations (after inversion) Ground surface temperature variations (after inversion) Ground surface temperature variations (after inversion)
Pasquale et al. (2000, 2005)
Glacial and Holocene to secular temperature trends Secular temperature trends Secular temperature trends Temperature change in the last 100–300 years.
Rajver et al. (1998)
Correia & Safanda (2001)
Rimi (2000)
Mediterranean Climate Variability over the Last Centuries: A Review
From 0 to 40 m 1850–present Non-tropical water depth in the corals, (Cladocora Mediterranean Sea caespitosa) Slovenia 1500–1980s/ 9 Boreholes1990s subsurface temperature profiles Boreholes30000 BP Sloveniasubsurface –1992 Ljutomer temperature profiles
61
62 Mediterranean Climate Variability
and corals and their distribution. The second part of this section provides an overview on lower resolved natural new land and marine proxies (boreholes, vermetids, non-tropical corals and deep sea corals).
Tree-Ring Information from the Iberian Peninsula and Italy Many past studies have described the use of tree-ring or dendroclimatic data to reconstruct past variations in precipitation, temperature, soil moisture, streamflow, the frequency of extreme droughts and atmospheric circulation indices. From dendroclimatic reconstructions over the Iberian Peninsula, several periods of differentiated climatic conditions have been highlighted over the last several hundred years (Creus-Novau et al., 1997; Saz and Creus-Novau, 1999; Saz, 2004). Creus-Novau et al. (1992), Saz et al. (2003) and Saz (2004) used treering information to reconstruct temperature and precipitation in different points of the northern half of Spain since the fifteenth and sixteenth centuries and over entire Spain for the last millennium. Several periods of differentiated climatic conditions have been highlighted over the last several hundred years (CreusNovau et al., 1997; Saz and Creus-Novau, 1999; Saz, 2004). Creus-Novau et al. (1992), Saz et al. (2003) and Saz (2004) used tree-ring information to reconstruct temperature and precipitation in different points of the northern half of Spain since the fifteenth and sixteenth centuries and over entire Spain for the last millennium. The results are shown in Fig. 14. They used a set of 42 dendrochronologies constructed from more than 1,500 samples of different trees
Figure 14: Annual mean temperature and total annual precipitation in northeastern Spain during the 1500–1900 period. Grey curve: standardized anomalies (reference period 1850–1900). The black curve is a 15-year running average (Saz, 2004).
Mediterranean Climate Variability over the Last Centuries: A Review
63
and from different tree species. More than 90% of the cores were extracted from coniferous trees. Climate reconstructions obtained from these chronologies allow studying the evolution of spring, summer, fall and winter (and annual) temperature and precipitation since the fifteenth century for nine different stations of Spain located in different bioclimatic areas. During the first centuries of the last millennium Iberian climate was characterized by high temperatures and precipitation, as well as by a remarkable climatic regularity that lasted till the mid-fourteenth century, when a shift in Iberian climate took place. This led to increased climate variability with a remarkable reduction in temperatures and an intensified occurrence of precipitation extremes. The LIA, which reached its maximum during the seventeenth century and lasted up to the early decades of the nineteenth century, was also manifested over the Iberian Peninsula as a period of cold conditions and increased climate variability, supported by lagoon and coastal sedimentary records (Luque and Julia´, 2002; Desprat et al., 2003). These cold phases coincide with similar periods described in western and central Europe. As for rainfall, the most important dry anomalies appear in the sixteenth and seventeenth centuries, a period with high interannual temperature and precipitation variability. Creus-Novau et al. (1995) used tree-ring information to reconstruct the climatic conditions in Galicia (northwestern Spain) for the last centuries. Galli et al. (1994) used Pinus pinea L. from Ravenna pine forest to check the possibility of reconstructing winter temperatures for the 1653–1985 period for an area close to the Adriatic coast, Italy. Briffa et al. (2001) used a large number of tree-ring data from southern Europe (Spain, Italy, southern Balkans, Greece) in order to reconstruct mean averaged central and southern European growing season (April–September) temperature series back to the early seventeenth century. Recently, Budillon et al. (2005), found both hyperpycnal flows from flood-prone stream and tempestites appearing as sand layers in the stratigraphic record of shelf areas are proxies for past storminess. The case study of the Salerno Bay shelf record from Central Italy revealed at least four events related to major storms that occurred in the area during the last 1,000 years (1954, 1879, 1544 and an older unknown event).
Tree-Ring Information from South Eastern Europe and Eastern Mediterranean Dendroclimatology in the eastern Mediterranean region is still in the early stages of development. Most studies are recent with the exception of a few earlier works. Gassner and Christiansen-Weniger (1942) demonstrated that tree growth is significantly influenced by precipitation in parts of Turkey. B. Bannister, from the Laboratory of Tree-Ring Research at The University of Arizona, was the first dendrochronologist to attempt systematic tree-ring
64 Mediterranean Climate Variability
dating of Near Eastern archaeological sites (Bannister, 1970). He collected and analysed tree-ring specimens from a 800 BC tomb in Turkey and carried out preliminary examinations of wood samples from Egyptian coffins. He also collected and cross-dated samples of Cedar of Lebanon (Cedrus libani) in Lebanon. Several dendrochronological studies have followed the work of Gassner and Christiansen-Weniger (1942) and Bannister (1970). For example, a large number of tree-ring chronologies, mainly in Greece and Turkey, for dating archaeological sites are produced by Kuniholm and Striker (1987) and Kuniholm (1990, 1994). During the past six years, dendroclimatology has begun to establish itself in the region through multi-national scientific projects that are interested in understanding climate variability over several centuries. The first dendroclimatic reconstruction (a 396-year-long reconstruction of October–May precipitation based on two chronologies of Juniperus phoenicia) in the Near East was developed by Touchan and Hughes (1999) and Touchan et al. (1999) in southern Jordan. They showed that the longest reconstructed drought, as defined by consecutive years below a threshold of 80% of the 1946–1995 mean observed October–May precipitation, lasted four years. More recently in Turkey, Akkemik (2000) investigated the response of a Pinus pinea tree-ring chronology from the Istanbul region to temperature and precipitation. Hughes et al. (2001) demonstrated that the cross-dating in archaeological specimens over large distances in Greece and Turkey has a clear climatological basis, with signature years consistently being associated with specific, persistent atmospheric circulation anomalies. D’Arrigo and Cullen (2001) presented the first 350-year (1628–1980) dendroclimatic reconstruction of February–August precipitation for central Turkey (Sivas), although it relied on Peter Kuniholm’s materials that end in 1980. Touchan et al. (2003) used tree-ring data from living trees in southwestern Turkey to reconstruct spring (May–June) precipitation several centuries back in time (Fig. 15). Their reconstructions show clear evidence of multi-year to decadal variations in spring precipitation. The longest period of spring drought was only four years (1476–1479). The longest reconstructed wet periods were found during the sixteenth and seventeenth centuries. They also found that spring drought (wetness) is connected with warm (cool) conditions and southwesterly (continental) circulation over the eastern Mediterranean. A subsequent reconstruction was developed by Akkemik et al. (2005) for a March–June precipitation season from oak trees in the western Black Sea region of Turkey. They found that during the past four centuries drought events in this region persisted for no more than two years. Akkemik and Aras (2005) reconstructed April–August precipitation (1689–1994) for the southern part of central Turkey region by using Pinus nigra tree rings. These various tree-ring studies in Turkey suggest that the duration of dry years generally extends for one or two years and rarely for more than
Mediterranean Climate Variability over the Last Centuries: A Review
65
Figure 15: Time series plot of reconstructed May–June precipitation, 1339–1998. Horizontal solid line is the mean of the observed data. Horizontal dashed line is the arbitrary threshold of 120% of the 1931–1998 mean May–June precipitation (80.33 mm). Horizontal dotted line is the arbitrary threshold of 80% of the 1931–1998 mean May–June precipitation (53.55 mm). The instrumental record is in grey. Touchan et al. (2003) Preliminary reconstructions of spring precipitation in southwestern Turkey from tree-ring width. Int. J. Climatol., 23, 157. Copyright ß (2005) the Royal Meteorological Society, first published by John Wiley & Sons Ltd.
three years. In accordance with other studies, the years 1693, 1725, 1819, 1868, 1878, 1887 and 1893, which were below two standard deviations from the twentieth century long-term mean, were determined as the driest years in the eastern Mediterranean basin (Akkemik and Aras, 2005). Touchan et al. (2005a) were the first to develop a standardized precipitation index (drought index) reconstruction from tree rings. Their study provided important regional information concerning hydroclimatic variability in the southwestern and south-central Turkey. Touchan et al. (2005b) continued their investigations of the relationships between large-scale atmospheric circulation and regional reconstructed May–August precipitation for the eastern Mediterranean region (Turkey, Syria, Lebanon, Cyprus and Greece). As part of this study, they conducted the first large-scale systematic dendroclimatic
66 Mediterranean Climate Variability
sampling for this region from different species. They developed six May–August reconstructions ranging in lengths from 115 to 600 years. The study found no long-term precipitation trends during the last few centuries. They also identified large-scale atmospheric circulation influences on regional May–August precipitation. For example, this precipitation season is driven by anomalous below (above) normal pressure at all atmospheric levels and by convection (subsidence) and small pressure gradients at sea level. A pioneering comparison of their tree-ring data and independent (i.e. sharing no common predictors in the reconstruction procedure) reconstructions of large-scale sea level pressure (SLP; Luterbacher et al., 2002b) and surface air temperature (SAT; Luterbacher et al., 2004) showed that large-scale climatic patterns associated with precipitation and tree-ring growth in this region have been substantially stable for the last 237 years. D’Arrigo and Cullen (2001), Akkemik et al. (2005) and Touchan et al. (2005a,b) have begun new investigations linking the dendroclimatic reconstructions to other proxy records, specifically to documentary evidence. These new studies provide examples of how historians and archaeologists can use dendroclimatic reconstructions to study and interpret the interactions between past human behaviour and the environment. For example, all four studies identified the year 1660 as dry summer while Purgstall (1983) reported that catastrophic fires and famine in Anatolia occurred in the same year.
Tree-Ring Information from Northern Africa Morocco has an interesting advantage in North Africa as the westerlies bring humid air towards the Rif and Atlas mountains, making possible the growth of millennium cedars on these mountains. Munaut (1982) sampled about 50 cedar sites (Cedrus atlantica), which are a source of important climatic information for that country (Till and Guiot, 1990). Figure 16 presents the
Year
Figure 16: Reconstruction of annual (October–September) precipitation (averaged over Morocco) by Till and Guiot (1990) from 46 cedar tree-ring series. The precipitation anomalies are expressed in departures from the period 1925–1970 in mm/year. In grey are the uncertainties for each year.
Mediterranean Climate Variability over the Last Centuries: A Review
67
annual precipitation for Morocco over the last around 1,000 years. It appears that the twentieth century was among the wettest of the last millennium. In comparison, the 1600–1900 period was 84 mm drier than the reference period (1925–1975), i.e. a deficit of about 11%. Meko (1985) and Chbouki et al. (1995) discussed temporal and spatial variation of Moroccan drought (based on tree-ring information). Till and Guiot (1990) published a 900-year reconstruction of October–September precipitation for three different areas in Morocco, indicating a continuous tendency towards a wetter climate during the twentieth century, and drier conditions than present during the sixteenth, seventeenth and eighteenth century. SerreBachet et al. (1992) pointed out the fact that the spatial variability of precipitation is difficult to be interpreted as some tree species used for reconstruction are related to winter conditions and others refer to the summer period. Nevertheless they showed that the dry climate periods reconstructed for Morocco and also for Spain reflect a climate out of phase with the rest of Europe, likely under a stronger effect of winter NAO than in eastern Mediterranean regions. Glueck and Stockton (2001) used climate-sensitive Moroccan tree-ring data (and ice-core data from Greenland) to reconstruct the winter North Atlantic Oscillation Index (NAOI) back to 1429. It is, however difficult to find a correlation between their NAOI and the precipitation series presented in Fig. 16 (Till and Guiot, 1990).
Speleothems Speleothems are secondary cave deposits, such as stalactites and stalagmites, formed when calcium carbonate (usually calcite) precipitates from degassing solutions seeping into limestone caves. Usually, most studies use stalagmites due to their simple geometry and rapid growth rate, which typically vary between approximately 0.05 and 0.4 mm/year. Provided that annual bands are present and well preserved, the combination of annual band counts and Uranium-series dating results in absolute chronologies with relatively small age uncertainties (Fleitmann et al., 2004). In addition, the thickness of annual bands can be used to reconstruct either temperature (Frisia et al., 2003) or amount of precipitation (e.g. Fleitmann et al., 2004), depending on the environmental settings in and above the cave. For instance, in regions with a predominantly arid to semi-arid climate, the thickness of annual bands is primarily controlled by the availability of water (Fleitmann et al., 2003). Oxygen (18O) and carbon (13C) isotopic ratios are currently the most frequently used stalagmite-based climate proxies. Both are capable to provide information on temperature and/or hydrological balance (Schwarcz, 1986; Baker et al., 1997; Bar-Matthews et al., 1997, 2000; Bar-Matthews and Ayalon, 2004; Ayalon et al., 1999; Desmarchelier et al., 2000; McDermott et al., 2001; Bard et al., 2002; Burns et al., 2002; Spo¨tl and Mangini,
68 Mediterranean Climate Variability
2002; Frisia et al., 2003, 2005; Kolodny et al., 2003; Fleitmann et al., 2004; Mangini et al., 2005). The use of 13C as climate proxy, however, has remained somewhat limited as it can be influenced by many, sometimes counteracting, parameters, which do not always relate to climate (Baker et al., 1997). Using conventional sampling techniques (e.g. a dental drill) temporal resolution of isotopic time-series typically ranges from 1 to 20 years. More recently, newly developed analytical techniques laser ablation inductively coupled mass spectrometry (LA-ICP-MS) allow the measurement of climate-sensitive trace elements (Mg, P, U, Sr, Ba and Na) at much higher (weekly to monthly) resolution (Baldini et al., 2002; Treble et al., 2003, 2005; Montagna, 2004). To date, few studies have focused on the reconstruction of continental climate variability during the past 500–1,000 years using speleothems, mainly due to the difficulty in devising high-resolution sampling strategies. The work of Frisia et al. (2003) reports on annual growth rates within single annual laminae in three contemporaneously deposited Holocene speleothems from Grotta di Ernesto (Alpine cave in northern Italy), which respond to changes in surface temperature rather than precipitation. Based on monitoring of present-day calcite growth, and correlation with instrumental data for surface climatic conditions, the authors interpreted a higher ratio of dark- to light-coloured calcite and the simultaneous thinning of annual laminae as indicative of colder-than-present winters. Such dark and thin laminae occur in those parts of the three stalagmites deposited from 1650 to 1713 and from 1798 to 1840, as reconstructed through lamina counts. An 11-year cyclicity in growth rate, coupled with reduced calcite deposition during historic minima of solar activity, suggests a solar influence on lamina thickness and temperature, respectively. Spectral analysis of the lamina thickness data also suggests that the NAO variability influenced winter temperatures. More recently, Antonioli et al. (2003) and Montagna (2004) examined a stalagmite collected from the Grotta Verde, located on Capo Caccia promontory on the northwest coast of Sardinia (Central Mediterranean Sea). The oxygen isotopic record by Antonioli et al. (2003) covers the last 1,000 years and reveals a centennial-scale variability with a resolution of 20 years. Their climate reconstruction clearly demonstrates the presence of a warm/wet ‘‘Medieval Warm Period’’, a cold/dry LIA and a warming trend since 1700. This rise in temperature ended around the years 1930–1940, and was followed by a relatively cold/dry period between the years 1940 and 1995. Based on these data, Montagna (2004) obtained a millennial-scale seasonally resolved record of precipitation variability from Sardinia. The 18O values show significant changes in precipitation during the last millennium, comparable with the lowfrequency signals observed in a long European tree-ring chronology (Esper et al., 2002). The speleothem 18O record between 1600 and 1800 indicates relatively
Mediterranean Climate Variability over the Last Centuries: A Review
69
dry and cold conditions corresponding to the LIA, followed by a gradual warmer and wetter trend, culminating in 1975. Moreover, alternating warm/wet and cold/dry conditions mark the period between 1000 and 1600. Very recently, Mangini et al. (2005) reconstructed the air temperature variation during the past two millennia using the 18O composition of a precisely dated stalagmite from the central Alps. Mangini and co-authors showed that the temperature maxima during the Medieval Warm Period (800–1300) were on average 1.7 C higher than the minima during the LIA. In addition, the Alpine stalagmite reveals a highly significant correlation with 14C, suggesting the importance of the solar forcing in the NH during the past two millennia.
Corals Apart from tree-ring information, it was shown that annually banded reef corals from the northernmost Red Sea (28 N–29.5 N) provide proxy records of temperature seasonality and interannual to multidecadal variations in temperature/aridity for the southeastern Mediterranean region (Egypt, Israel, Palestine, Jordan) during the past centuries, the Holocene epoch and the last interglacial period (Felis et al., 2000, 2004; Rimbu et al., 2001, 2003a). Isotopic and elemental tracers, incorporated into the carbonate skeletons of these massive corals during growth, provide proxies of past environmental variability of the surface ocean, reflecting variations in Sea Surface Temperature (SST) and hydrologic balance (e.g. Felis and Pa¨tzold, 2004). In the first step, an oxygen isotope record covering the past 250 years derived from a living coral (Fig. 17) was compared to instrumental records of gridded SST and land station precipitation from the region (Felis et al., 2000). However, the coral record was not calibrated against a single parameter to provide a quantitative reconstruction because both, the temperature and the hydrologic balance at the sea surface, influence coral oxygen isotopes. In the second step, the coral oxygen isotope record was compared with indices of the Arctic Oscillation (AO)/NAO (Rimbu et al., 2001). In a third step, the coral record was compared to fields of SST and surface wind in the eastern Mediterranean/Middle East region, in order to reveal the physical mechanism for the linkage between the AO/NAO and variations of SST and hydrologic balance in the northernmost Red Sea. This combined analysis of the proxy record and instrumental climate data revealed that the region’s interannual to decadal climate variability is controlled by a high-pressure anomaly over the Mediterranean Sea that is associated with the AO/NAO, especially during the winter season. This high-pressure anomaly favours an anticyclonic flow of surface winds over the eastern Mediterranean, thereby controlling the advection of relatively cold air from southeastern Europe towards the northern Red Sea (Rimbu et al., 2001). Enhanced variance
70 Mediterranean Climate Variability
Figure 17: Coral oxygen isotope record (1750–2000) from Ras Umm Sidd (northernmost Red Sea, 28 N) near the southern tip of the Sinai Peninsula (Felis et al., 2000). It shows detrended and normalized mean annual time series, based on bimonthly resolution data. On interannual to multidecadal timescales, coral oxygen isotopes reflect variations in temperature/aridity. Pronounced cold/ arid conditions occur during the 1830s. Note that on interannual to decadal timescales colder/more arid conditions in the northernmost Red Sea are associated with increased precipitation along the southeastern margin of the Mediterranean Sea and decreased precipitation in Turkey. The bold line represents 3-year running average.
at interannual periods of 5–6 years was observed in all coral records from the northernmost Red Sea, and was also detected in a tree-ring-based reconstruction of Turkish precipitation covering the 1628–1980 period (D’Arrigo and Cullen, 2001). It was identified as a stable feature of eastern Mediterranean/Middle East climate, and it was shown to be characteristic for the influence of the AO/NAO on the region’s climate variability over longer periods (Felis et al., 2000, 2004). Further prominent oscillations identified in the coral-based climate reconstruction of the past 250 years from the northernmost Red Sea have periods of about 70, 22–23 and 8–9 years (Felis et al., 2000). The latter two periods were also identified in the tree-ring-based precipitation reconstruction from Turkey (D’Arrigo and Cullen, 2001). To summarize, annually banded reef corals from the northernmost Red Sea provide a seasonally resolved archive for temperature and aridity variations in the southeastern Mediterranean (Egypt, Israel, Palestine, Jordan) and the influence of the AO/NAO on the region’s interannual climate variability during the past centuries, the Holocene epoch and the last interglacial period (Felis et al., 2000, 2004; Rimbu et al., 2001, 2003a).
Mediterranean Climate Variability over the Last Centuries: A Review
71
Evidence from Lower Resolution Natural Proxies and New Marine Archives This part provides a short overview on natural land and marine proxies including boreholes, vermetids, non-tropical corals and deep sea corals. Subsurface temperature information in the Mediterranean from borehole data Temperature-depth profiles measured in boreholes contain a record of temperature changes at the Earth’s surface. The geothermal method of past climate reconstruction based on the information recorded in temperature logs has become well established within the last decade (Lachenbruch and Marshal, 1986; Shen and Beck, 1991; Beltrami and Mareschal, 1995; Huang et al., 2000; Harris and Chapman, 2001; Beltrami and Bourlon, 2004; Pollack and Smerdon, 2004). The basic assumption of this method is that climate changes are accompanied by long-term temperature changes of the Earth’s surface, which propagate downwards by heat conduction and can be reconstructed as ground surface temperature histories. In the absence of moving fluids, changes in ground surface temperature (GST) diffuse slowly downwards and are manifested at a later time as anomalies in the Earth’s background temperature regime. Because the thermal diffusivity of rock is relatively low (10 6 m2 s 1), the effect of short-period variations like the annual changes of surface temperature will be measurable at a depth of a few tens of metres (depending on the specific thermal properties of the rock). Long-period events like the temperature changes associated with post-glacial warming will still produce significant signal amplitudes at much larger depths of around 1,000 m. However, each method of paleoclimatic reconstruction has strengths and limitations, and the reconstruction of past climatic changes from geothermal data is no exception to this rule. A paleoclimatic reconstruction derived from borehole temperatures is characterized by a decrease in resolution as a function of depth and time. Typically, individual climatic events can be resolved from borehole temperatures only if its duration was at least 60% of the time since its occurrence. i.e. an event that occurred 500 years before present will be seen as a single event if it lasted for at least 300 years; otherwise events are incorporated into the long-term average temperature change (Beltrami and Mareschal, 1995b). This decrease in resolution is due to heat diffusion and it is a real physical limitation of the method, which is not possible to overcome with mathematical tools. The resolving power of a borehole-based reconstruction is not only restricted by the diffusive nature of heat transfer, but also dependent on the level of non-climatic perturbations which affected the site as well as uncertainties in methodological aspects (Mann et al., 2003; Pollack and Smerdon, 2004). Surface factors such as changes in vegetation cover (Nitoiu and Beltrami, 2005), underground hydrology (Kohl, 1997; Reiter, 2005), and latent heat effects (Mottaghy and Rath, 2005), among other, can affect the underground thermal regime independently
72 Mediterranean Climate Variability
of climate. Therefore, borehole data must be screened carefully before they are analysed for climate signatures. Because of decreasing resolution of the method for older events, its most promising application is the GST reconstruction of the last few centuries. The subsurface climatic signal of this period is contained in the uppermost 150–200 m of the temperature profiles, but the minimum borehole needed for analysis is about 300 m to allow for a robust estimation of the quasi steady-state profile (Hartmann and Rath, 2005). In many European and North American sites, the GST history of the last two centuries obtained with this method can be compared directly with the observed surface air temperature series measured at nearby meteorological stations. For earlier periods, the geothermal method offers an alternative source of information to infer temperature evolution through the last centuries which can be potentially compared with proxy records at regional and hemispheric/global scales (Beltrami et al., 1995; Rajver et al., 1998; Beltrami, 2002; Briffa and Osborn, 2002). The international heat flow community has supported the creation of a database of borehole temperature logs as an archive of geothermal signals of climate change (Huang and Pollack, 1998). Currently the database contains over 800 borehole reconstructions and most of the original borehole temperature profiles; over 75% of these data are located in the NH. The geographic coverage is densest in North America and Europe, with substantial datasets from Asia. In the Mediterranean area, the coverage of temperature logs is relatively sparse. Nevertheless, some studies have focused in reconstructing recent climate trends in different parts of the Mediterranean area with the purpose of identifying potential secular warming/cooling in geothermal information and comparing it to available meteorological and proxy evidence. In the central Mediterranean area Rajver et al. (1998) analysed nine boreholes from Slovenia and obtain a warming of 0.7 C for the last century, which is found to be compatible with meteorological observations. The Ljubljana borehole (1,965 m depth) allowed for a 90 Ka reconstruction, which was favourably compared with paleorecords of temperature for neighbouring regions (Hungary and the Alps). Pasquale et al. (2000, 2005) analyse boreholes in the western and eastern parts of the Apennines (Italy). They identify differences in the Tyrrhenian and Adriatic sides which are in correspondence with nearby meteorological observatories and support influence of local-scale microclimates, the western side showing a clear warming through the last century in the four analysed boreholes and the Adriatic side presenting clear cooling trends in correspondence with the 1940s to 1980s cold relapse in meteorological observations. In the western Mediterranean area, Correia and Safanda (2001) analyse subsurface temperatures in the Atlantic margins, i.e. a borehole close to Evora in Portugal showing a temperature increase of 1 C since the end of the nineteenth
Mediterranean Climate Variability over the Last Centuries: A Review
73
century. Rimi (2000) analysed 10 boreholes in four different climatic types in the north of Morocco. Results show varying warming amplitudes between 3 and 4 C, which are in some areas amplified by deforestation and mining effects. The above studies present the potential of subsurface temperature information as an alternative approach to establish the magnitude of secular temperature change in different Mediterranean areas. Proliferation of such studies to cover the larger Mediterranean area would be desirable. Such estimates should be compared with climate proxy reconstructions and model simulations to assess consistency and increase the robustness of our knowledge of past climate variations (Gonza´lez-Rouco et al., 2003a, 2006; Beltrami et al., 2005). In Mediterranean areas where the meteorological records are relatively short, borehole reconstructions can increase our perspective of temperature changes in the last centuries. Alternatively, the borehole approach can help to elucidate long-term temperature trends in cases where meteorological observations present potential inhomogeneities. Mediterranean sea temperatures and sea-water chemistry derived from Vermetids, non-tropical, and deep-sea corals The oceans exert a very strong influence on the atmosphere due to the continuous exchange of heat and water vapour with the atmosphere and they play a critical role in the chemical balance of the atmospheric system. SST is one of the most important variables for the Earth’s climate system. Changes in SST and their interactions with the atmosphere have the potential to affect the precipitation patterns, causing droughts, storms and other extreme weather events, mainly in regions particularly sensitive and potentially very vulnerable to climate changes, such as the Mediterranean region. Thus, the possibility to derive long high-resolution time series of key climatic parameters such as SST and salinity for the Mediterranean Sea is a fundamental prerequisite for a better understanding of the mechanisms governing climate change in this region. The only possibility to extend our climate database far beyond the instrumental record is studying the elemental and isotopic composition of welldated natural archives of climate variability. Over the last centuries, SST records of high resolution (annual to seasonal) have not been available in the temperate area of the Mediterranean Sea due to the absence of appropriate proxies, such as the corals in tropical and sub-tropical seas. An exception are the annually banded reef corals of the northernmost Red Sea that provide a seasonally resolved archive of past climate variability for the southeastern Mediterranean region. With this in mind, great attention has been paid in recent years to obtain high-resolution records of SST, salinity and water chemistry for the Holocene in the Mediterranean Sea, using new archives such as vermetid reefs (living reefs, last 500–600 years, 30–50 years resolution), non-tropical corals (i.e. living
74 Mediterranean Climate Variability
Cladocora caespitosa, last 100–150 years, seasonal to weekly resolution), and deep-sea corals (i.e. living Desmophyllum dianthus, Lophelia pertusa, Madrepora oculata, last 100–150 years, annual to seasonal resolution). These new archives will complement and improve the information derived from the main climate indicators such as foraminiferal tests, alkenones, dinoflagellate cysts, calcareous nanoplankton, and, especially for the Mediterranean Sea, serpulid overgrowth on submerged speleothems (Antonioli et al., 2001). All these latter marine markers enable longer paleoclimate reconstructions but with a much coarser resolution (usually lower than 100–200 years). However, in areas of extremely high sedimentation rates (>80 cm/ka), such as in the distal part of the nilotic cell, southern Levantine Basin, foraminiferal stable isotope composition clearly show the evidence of both the Medieval Warm Period and the LIA, with a resolution of 40–50 years during the last millennium (Schilman et al., 2001). Vermetids are thermophile and sessile gastropods living in intertidal or shallow subtidal zones, forming dense aggregates of colonial individuals. Vermetids show a wide areal distribution, also being present in the Mediterranean Sea, such as in Syria, Lebanon, Greece, Turkey, Crete, Italy etc. (Safriel, 1966, 1974; Pirazzoli and Montaggioni, 1989; Pirazzoli et al., 1996; Delongeville et al., 1993; Antonioli et al., 1999, 2001; Silenzi et al., 2004; Fig. 18). Vermetus triquetrus (Bivona-Bernardi, 1832) and Dendropoma petraeum (Monterosato, 1892) are the two species forming clusters in the Mediterranean Sea. Presently, living reefs in the northwestern coast of Sicily present a fossil
Figure 18: Vermetid reefs world distribution (black dots). The vast majority of species is located in several world seas and oceans, between 44 N and 44 S (Silenzi et al., 2004).
Mediterranean Climate Variability over the Last Centuries: A Review
75
portion that is maximum about 650-years old, whereas some fossil samples collected in uncemented tempestite deposits date back to 2500 years BP. Vermetids are generally used as indicators of sea level changes and neotectonic stability (Stephenson and Stephenson, 1954; van Andel and Laborel, 1964; Kemp and Laborel, 1968; Hadfield et al., 1972; Laborel and Delibrias, 1976; Focke, 1977; Jones and Hunter, 1995; Angulo et al., 1999; Antonioli et al., 1999, 2002) due to the possibility of precisely dating their calcite skeleton by radiometric methods. Silenzi et al. (2004) analysed and compared two sections of Dendropoma sp., from NW Sicily, spanning 500 years to the present day (Fig. 19).
Figure 19: Left: Vermetid reef mushroom-like type near Capo Gallo promontory (NW Sicily) and Right: Vermetid reef platform type near S. Vito lo Capo (from Silenzi et al., 2004).
The spatial resolution between two neighbouring oxygen data corresponds to 30–50-years time interval. The isotopic records show a clear oscillation, with 18O values being more positive than at present during the period between the years 1600 and 1850. Data indicate a maximum difference in 18O from the LIA to present day of about 0.38 0.1ø, which corresponds to 1.99 0.37 C SST difference. After the LIA, vermetid reefs recorded the warming trend that characterized the last century. This rise in temperature ended around the years 1930–1940, and was followed by a relatively cold period until 1995. Moreover, the SST reconstruction clearly demonstrates that in the early to mid-1500s, SSTs were warmer than today. The study by Silenzi et al. (2004) proved that vermetid reefs have the potential to be excellent indicators of SST variability both in historical time (actual growing reef) and during the Holocene (fossil reefs), allowing paleoclimatic reconstructions at high temporal resolution. Shallow water scleractinian corals secrete a calcareous skeleton whose minor and trace element composition provides a potentially unique archive of the ambient environment in which it grew (e.g. Felis and Pa¨tzold, 2004). To date, there have been few studies dealing with coral chemistry at high latitudes but
76 Mediterranean Climate Variability
none investigated coral species in the Mediterranean Sea. Recently, Montagna et al. (2004) and Silenzi et al. (2005) targeted a temperate coral species living in the Mediterranean Sea with promising results. This non-tropical coral deposits two bands per year, one high-density band forms during periods of low temperatures and low light intensity (autumn–winter), whereas the low-density band corresponds to high temperature and high light intensity (spring–summer). Silenzi et al. (2005) demonstrated for the first time the feasibility of using C. caespitosa as a paleoclimate archive of SST, through isotopic (18O and 13C) and elemental (Sr/Ca and Mg/Ca) analyses on a 95-year corallite collected along the continental shelf of the Ligurian Sea. The work by Montagna et al. (2004) further proved the potentiality of C. caespitosa as a paleothermometer. Geochemical ratios (Sr/Ca and B/Ca) exhibit a close relationship to the in situ measured (weekly) SST at the sampling site and in particular, B/Ca shows an extraordinary high degree of correlation (r ¼ 0.88, n ¼ 130) with SST. Both the studies thus demonstrate the capability of this long-lived (100–150 years) non-tropical coral to preserve SST changes in the Mediterranean Sea at seasonal to weekly resolution over the last 100 years. In addition, the combination of 18O and trace element ratios (Sr/Ca, B/Ca) will allow to track past salinity changes at the same resolution of SST. The use of the calibration equations obtained from the Ligurian and the Adriatic Sea will enable the reconstruction of the paleo-SSTs during periods particularly interesting for climate studies. Deep-sea corals are potentially excellent archives of past oceanographic conditions with their wide depth range providing climate proxies for intermediate and deep waters. Deep-sea corals can be directly dated using high precision 234 U/238U/230Th, 226Ra/210Pb and 14C methods (Cheng et al., 2000; Goldstein et al., 2001; Adkins et al., 2002, 2004; Frank et al., 2004; Pons-Branchu et al., 2005). An understanding of the physical and chemical parameters of deep waters is important for climate reconstructions since atmospheric climatic conditions are intimately linked or coupled with the ocean circulation patterns. The great potential of these archives in the deep-water realm stems from the fact that they can provide higher resolution than sediment cores and they are not affected by bioturbation. Moreover, they can span more than 100 years, allowing obtaining decadal changes with seasonal resolution. Fossil deep-sea corals, such as Lophelia pertusa, Madrepora oculata and Desmophyllum dianthus, have been widely documented in the Mediterranean basin whereas living specimens seem to be less common and widespread (Taviani et al., 2005, and references therein). Montagna et al. (2005a) studied the trace and minor element compositions in two Desmophyllum dianthus specimens collected in the Mediterranean Sea and in the Great Australian Bight. The chemical variation of productivity controlled elements, such as P, Mn and Ba seems to reflect changes in seawater
Mediterranean Climate Variability over the Last Centuries: A Review
77
concentrations, demonstrating the potentiality of this species to become a powerful archive of deep water chemistry. In addition, the typical watertemperature sensitive elements are the subject of ongoing research, which aims to reconstruct the temperature variations at high resolution in the deep-water realm (Montagna et al., 2005b). These results illustrated the potential use of geochemical composition in deep-sea corals within the Mediterranean Sea, providing an important new approach to help unravel climatic variability with respect to the temporal evolution of the chemical and physical properties of intermediate and deep waters.
1.3. Large-Scale Climate Reconstructions and Importance of Proxy Data for the Mediterranean The first part of this section shortly describes the basic idea how to incorporate climate information from different areas in a statistical way to reconstruct largescale climate fields. The second part addresses the importance of documentary and natural proxies for Mediterranean precipitation and temperature field reconstructions at seasonal timescales. Different from local or regional climate reconstructions using a variety of documentary or natural proxies (Sections 1.1 and 1.2), multivariate statistical climate field reconstruction (CFR) techniques include multivariate calibration of proxy data against instrumental records. CFR seeks to reconstruct a large-scale field, such as surface air temperature, pressure or precipitation using a spatial network of proxy indicators, performing a multivariate calibration of the large-scale information in the proxy data network against the available instrumental data. This so-called ‘‘upscaling’’ involves fitting statistical models, which are mostly regression-based, between the local proxy data and the large-scale climate. Model fitting is usually based on the overlap period between proxy and instrumental data. It is assumed that the statistical relationships throughout the reconstruction period are stable (concept of stationarity). Because the large-scale field is simultaneously calibrated against the entire information in the network, there is no a priori local relationship assumed between proxy indicator and climatic variable. All indicators should, however, respond to some aspect of local climate during part of the year (e.g. Guiot, 1992; Mann et al., 1998, 1999; Briffa et al., 2002; Jones and Mann, 2004; Luterbacher et al., 2004; Casty et al., 2005a,b; Guiot et al., 2005; Pauling et al., 2005; Rutherford et al., 2005; Xoplaki et al., 2005). The approach of CFR (e.g. Mann et al., 1998, 2000, 2005; Briffa et al., 2002; Luterbacher et al., 2002b, 2004; Mann and Rutherford, 2002; Bro¨nnimann and Luterbacher, 2004; Casty et al.,
78 Mediterranean Climate Variability
2005a,b; Pauling et al., 2005; Rutherford et al., 2005; Xoplaki et al., 2005; Zhang et al., 2005) provides a distinct advantage over averaged climate reconstructions for instance, when information on the spatial response to external forcing (e.g. volcanic, solar) is sought (e.g. Shindell et al., 2001a,b, 2003, 2004; Fischer et al., 2005). Thus, CFR allows insight into both the spatial and temporal details about past climate variations over the Mediterranean region. There are a few CFR reconstructions available covering the Mediterranean area: Guiot (1992) used a combination of documentary proxy evidence and natural proxies (tree rings, ice core data) to provide fields of annual temperature estimates from 1068 to 1979 for Europe, including a large part of the Mediterranean area. He found significant connection between northwest Europe and the central Mediterranean region during the LIA, while the western Mediterranean region had not experienced any significant cooling. Briffa et al. (2002) used treering maximum latewood density data to reconstruct large-scale patterns of warmseason (April–September) mean temperature for the period 1600–1887 for the NH, including the Mediterranean. Mann (2002b) used proxy data and long documentary and instrumental records to reconstruct and interpret large-scale surface temperature patterns back to the mid-eighteenth century for the Middle and Near East. This study suggested that interannual temperature variability in these regions in past centuries appears to be closely tied to changes in the NAO. Luterbacher and Xoplaki (2003) provided a preliminary 500-year long winter mean precipitation and temperature time series over the larger Mediterranean land area, calibrated in the twentieth century and reconstructed from long instrumental station series, documentary evidence and a few tree-ring data. They report also on the spatial temperature anomaly distribution for cold and warm winter extremes. The first question, however, is which proxies are of relevance for temperature and precipitation CFR at seasonal timescale. Pauling et al. (2003) investigated the importance of natural and documentary proxies for seasonal European and North Atlantic temperature field reconstructions. Using a set of 27 annually resolved proxies, they employed backward elimination techniques (e.g. Ryan, 1997) to identify the most important predictor at each gridpoint. These analyses included tree rings, ice core parameters, corals, a speleothem and indices based on documentary data. For boreal winter (October–March) they found the speleothem from Scotland and tree rings to be the most important proxy for the western Mediterranean; documentary evidence for the northern basin and parts of northern Africa, whereas the Red Sea coral is of relevance for the eastern basin. For boreal warm-season (April–September) temperatures, tree rings, documentary data and the speleothem proved to be most important. The importance of the speleothem is particularly striking as only one single series
Mediterranean Climate Variability over the Last Centuries: A Review
79
was used in the backward elimination analysis while other proxy types were more numerous. We performed a similar evaluation of high-resolution natural and documentary proxies for precipitation field reconstructions, though restricting the analyses over the larger Mediterranean land areas (10 W–40 E; 30 N–47 N). We followed the same method as described by Pauling et al. (2003). Since different proxies may record climate conditions at different times of the year, we study the performance of the proxy information for both the boreal cold (October–March) and warm (April–September) season (Pauling et al., 2003). Figures 20 and 21 (upper panels) depict the locations of the proxies used in this experiment (tree rings, ice cores, one coral, one speleothem, and several precipitation indices based on documentary data). Those proxies have been shown to significantly respond to local and regional precipitation during both the boreal cold and warm season within the twentieth century (not shown). Unfortunately, there are not many proxies fulfilling those criteria and only a few stem from the Mediterranean area. Most of the natural proxies have been downloaded from the NOAA World Data Center for Paleoclimatology, Boulder, Colorado, USA (http://www.ngdc.noaa.gov/wdc/wdcmain.html). As documentary indices are not available for the twentieth century, seasonally resolved indices based on instrumental measurements were degraded using a similar approach as Mann and Rutherford (2002), Pauling et al. (2003) and Xoplaki et al. (2005). Normally distributed white noise was added to the series to ensure the resulting pseudo-documentary indices are of similar quality as documentary indices derived from documentary evidence. Most of the documentary and natural proxies are available for a few centuries and can, thus, be potentially used for precipitation reconstructions (see below). The common period (1902–1983) of both the predictors and the predictands was used for calibration. First, for each gridpoint multiple regression models were established. Second, all but one predictor at each gridpoint was eliminated using backward elimination techniques. The last predictor at each grid point is regarded as the most important one of the initial predictor set (Pauling et al., 2003). Figure 20 (middle panel) presents the spatial distribution of the last remaining predictor at each gridpoint for the boreal cold season, derived through backward elimination. The pseudo-documentary indices are the most important predictors over large parts of continental Europe, explaining moderate 10–20% of the variance (Fig. 20, lower panel). There are larger areas in northern Africa, southeastern Europe and the Near East where tree-ring data are the most important proxy during boreal winter (Fig. 20, middle panel). The Scottish speleothem, the Red Sea coral and the ice core data from Greenland do only indicate small regions where these natural proxies are of major importance. Hence, the speleothem is
Figure 20: Top: Locations of the natural and documentary proxies used in this study. Middle: Spatial distribution of the most important predictors for boreal winter (October– March) precipitation over the larger Mediterranean area. For the legend see top panel. Bottom: Correlation of the most important predictor (Fig. 20, middle panel) with local precipitation over the period 1902–1983. All values above 0.21 are statistically significant at the 5% level (t test). The gridded precipitation dataset (10 W–40 E; 30 N–47 N; 0.5 0.5 resolution) from Mitchell et al. (2004) and Mitchell and Jones (2005) was chosen as the dependent variable. As different proxies have been used to produce this correlation map, the correlation sign changes over small spatial scales. These changes often correspond to changes of the most important predictor (Fig. 20, middle panel).
Mediterranean Climate Variability over the Last Centuries: A Review
81
much less important for precipitation than for temperature, as described by Pauling et al. (2003). The physical interpretation for those statistically resolved patterns is not trivial and outside the scope of this contribution. Concerning boreal summer, documentary precipitation series are still the most important proxies for large parts of the northern Mediterranean area and regions over the Iberian Peninsula as well parts of Morocco (Fig. 21). As for the boreal cold season, those proxies can account for a maximum of around 20% of summer precipitation over these areas. Compared with the cold season, tree rings have taken the place of the pseudo-documentary precipitation indices over southeastern Europe, Turkey and the Near East. These findings are in agreement with the results presented in Section 1.2.2. Tree rings explain between 10 and 20% of the April–September precipitation variability over these areas. Further, tree rings cover a large area of the Iberian Peninsula and northern Africa. Surprisingly, ice core data from Greenland seem to be of importance over the southern Near East pointing to possible teleconnections during the twentieth century. It has to be stressed that these findings are only meaningful where summer precipitation regularly occurs and is high enough to exhibit some variability. Over the southern part of the study area this is not the case or only for the late spring months/early autumn. Therefore, the importance of tree rings for boreal summer precipitation reconstructions over North Africa is clearly limited (Fig. 21). Further analyses have to prove, whether such relations derived within the twentieth century are stable back in time including more systematic testing of a larger dataset of proxies. It should be also taken into account that not only the proxy type determines the results of this preliminary analysis but also its initial number and location. Pauling et al. (2003) used different proxy types situated in the vicinity of each other, which reduces the influence of the location and allows the proxy characteristics to compete in the backward elimination process. These findings, though, have to be further examined for the larger Mediterranean area as well. Mann et al. (1998, 2000); Mann (2002a); Pauling et al. (2003, 2005); Guiot et al. (2005); Rutherford et al. (2005); Xoplaki et al. (2005) and the results presented above point out that the multi-proxy approach exploits the complementary strengths from each of the proxies to estimate temperature and precipitation change over a large area back in time. Thus, large-scale climate reconstructions based on a careful selection of a combination of temperature–precipitation-sensitive proxies from all over Europe, including the Mediterranean, provides a reliable means for reconstructing past regional and seasonal climate variability.
82 Mediterranean Climate Variability
Figure 21: As Fig. 20, but for boreal summer (April–September) precipitation.
Mediterranean Climate Variability over the Last Centuries: A Review
83
1.4. Mediterranean Winter Temperature and Precipitation Variability over the Last 500 Years This section presents the evolution of Mediterranean temperature and precipitation over the last 500 years using data presented in the Sections 1.1–1.3. The reconstructions are based on principal component, multivariate regression and have been extensively calibrated within the twentieth century. The reconstruction of entire temperature and precipitation fields, using multivariate calibration of the proxy data against the instrumental records, allow both spatial and temporal considerations about past climate variability over the Mediterranean (Section 1.3). An estimate of the Mediterranean mean temperature or precipitation can, for instance, be derived by averaging over the reconstructed patterns (see below). Information regarding the underlying spatial pattern (such as the different Mediterranean sub-areas) is, however, retained (Mann, 2002a). We will further present spatial fields of the coldest and mildest as well as the wettest and driest Mediterranean winters derived from the reconstruction period. Further, we also provide anomaly winter temperature and precipitation composites where we highlight the difference between multidecades of mild (wet) minus cold (dry) Mediterranean winters. Using a few natural proxies in combination with documentary data presented in Section 1.3 (Figs. 20, 21) and long instrumental station series we fit a statistical model to the winter (December–February average) Mediterranean temperature and precipitation for the land areas 10 W to 40 E and 35 N to 47 N. The details on the methodology and data used can be found in Luterbacher et al. (2004) for temperature and Pauling et al. (2005) in case of precipitation. Apart from the description and interpretation in terms of trends and uncertainties, we will also perform a wavelet analysis and will report on the change of distribution of winter extremes over the last 500 years. In addition, a PDSI is derived from these data for selected areas (Morocco, Italy and Greece). Figures 22 and 25 show the averaged winter mean Mediterranean temperature and precipitation anomalies (with respect to 1961–1990) from 1500 to 2002. The time series are composed of a reconstructed time period between 1500 and 1900 as well as the gridded Mitchell et al. (2004) and Mitchell and Jones (2005) data for the period 1901–2002. Figures 22 and 25 also present 30-year smoothed time series employing boundary constraint optimized to resolve the non-stationary late (end of the twentieth century, beginning of the twenty-first century) behaviour of the time series (Mann, 2004). As proposed by Mann (2004) we employ an objective measure of the quality of fit (mean-squared error, MSE) of a 30-year smooth with respect to the original time series.
84 Mediterranean Climate Variability
Figure 22: Winter (DJF) averaged-mean Mediterranean temperature anomalies (with respect to 1961–1990) from 1500 to 2002, defined as the average over the land area 10 W to 40 E and 35 N to 47 N (thin black line). The values for the period 1500–1900 are reconstructions; data from 1901 to 2002 are derived from Mitchell et al. (2004), Mitchell and Jones (2005). The thick black line is a 30-year smooth ‘‘minimum rough’’ constraint (mean squared error, MSE ¼ 0.866) calculated according to Mann (2004). The dashed horizontal lines are the 2 standard deviations of the period 1961–1990. The warmest and the coldest Mediterranean winters for the reconstruction and the full period are denoted.
The reconstructed winter near-surface air temperature (Fig. 22) time series is more stationary than the one for precipitation (Fig. 25), especially with respect to the amount of interannual variability. Strong departures from the 1961–1990 irregularly occurred during the entire 500-year period, although warm anomalies appeared to be enhanced during the beginning of the seventeenth century (with 1606–07 being the warmest winter within the reconstruction period) and during the twentieth century. There is a substantial warming trend starting around 1890. Centennial temperature variability increases after 1800. However, the uncertainties (two standard errors based on unresolved variance within the twentieth century calibration period, not shown) associated with the averaged winter temperature reconstructions are of the order of 1.1 C for single winters up to the 1660s, and decrease to around 0.5 C at the end of the nineteenth century.
Mediterranean Climate Variability over the Last Centuries: A Review
85
Note that filtered uncertainties of both, Mediterranean averaged near-surface air temperature and precipitation are presented in Section 1.7 (Figs. 43, 44). The warmest (coldest) winter was 1955 (1891). A glance at the most recent three winters (2002–03; 2003–04; 2004–05; not shown) which are based on the gridded analysis of Hansen et al. (2001) reveal, that except for 2003–04 (0.32 C warmer than 1961–1990), the two other winters were below the 1961–1990 reference period ( 0.28 C for 2002–03 and 0.44 C for 2004–05). The spatial anomaly of the coldest (1890–91; 2.4 C colder than the 1961–1990 reference period) and warmest winter (1606–07; 1.4 C warmer compared with the 1961–1990 period) for the reconstruction period (i.e. 1500–1900) are presented in Fig. 23. The anomaly spatial temperature maps of both winters shows a monopole pattern with above (1606–07) and below (1890–91) temperature anomalies all over the Mediterranean area. The largest deviations are found generally north of around 41 N. In the case of 1606–07, the uncertainties of the reconstructions are rather large (of the order of 1 C averaged over the entire area, smaller in the northern part, larger in the southern and eastern regions) as no instrumental data are available from the area at that time. The reconstructions are based on a few temperature indices derived from documentary evidence from central and eastern Europe and Western Baltic Sea conditions (Koslowski and Glaser, 1999) and an ice core derived temperature from Greenland (Vinther et al., 2003a; see Luterbacher et al., 2004, supplementary online material for the used climate information). Thus, these spatial reconstructions of earlier centuries should be treated with caution. The comparison with the absolutely mildest winter (1954–55) reveals also a monopole pattern, though the largest deviations are found over the southeastern Mediterranean (not shown). The reconstruction of the cold winter 1890–91 is much more reliable as there are high-quality instrumental data available from the Mediterranean area as well. Uncertainties are largest along the northern African coast and the southeastern basin (not shown). The warmest (coldest) decade was 1993–2002 (1680–1689) whereas the warmest (coldest) 30 Mediterranean winters in a row were experienced from 1973 to 2002 (1880–1909) with 0.16 C ( 0.85 C) departures from the 1961–1990 average. The anomalous spatial temperature distribution of the warmest and coldest 30 winters is presented in Fig. 24. Except for a few single gridpoints, all regions around the Mediterranean area experienced negative temperature anomalies during the 1880–1909 cold period compared with the 1961–1990 mean. For the warmest 30 winters (Fig. 24, bottom) from 1973 to 2002 the most positive departures are found over the northern areas and northwestern Africa, whereas the southeastern Mediterranean area experienced colder winters compared with the 1961–1990 average.
86 Mediterranean Climate Variability
Figure 23: Anomalous surface air temperature for the warm Mediterranean winter 1606–07 (top) and the cold Mediterranean winter 1890–91 (bottom) (with respect to 1961–1990).
Fischer et al. (2006), analysed the European climatic response to 16 major tropical eruptions over the last half millennium. They found winter cooling (though significant only over parts of the Iberian Peninsula and northern Africa) over the entire Mediterranean during the first post-eruption year (in contrast to northern Europe where a winter warming is experienced). The anomaly pattern for the second winter after the eruptions reveals a different pattern with a slight cooling in the western and eastern part and a warming in the remaining parts of the Mediterranean. These anomalies though, are not significant.
Mediterranean Climate Variability over the Last Centuries: A Review
87
Figure 24: Anomalous winter (DJF) temperature composites. Top: Averagedmean Mediterranean land surface air temperature for the 30 coldest winters in a row (1880–1909) over the last 500 years minus the 1961–1990 reference period (in C). Bottom: As top, but for the 30 warmest winters (1973–2002 minus 1961–1990). Data from 1880–1900 are reconstructions, data from the twentieth/twenty-first century stem from Mitchell et al. (2004) and Mitchell and Jones (2005). The averaged Mediterranean winter precipitation series (Fig. 25) clearly indicates reduced variability prior to around 1780, possibly an indication of a low number of proxy information available (Casty et al., 2005b; Pauling et al., 2005). Low-frequency variations also tend to rise over the centuries. The uncertainties (two standard errors, not shown) associated with the averaged winter
88 Mediterranean Climate Variability
Figure 25: Winter (DJF) averaged-mean Mediterranean precipitation anomalies (with respect to 1961–1990) from 1500 to 2002, defined as the average over the land area 10 W to 40 E and 35 N to 47 N (thin black line). The values for the period 1500–1900 are reconstructions (Pauling et al., 2005); data from 1901 to 2002 are derived from Mitchell et al. (2004) and Mitchell and Jones (2005). The thick black line is a 30-year smooth ‘‘minimum slope’’ constraint (mean squared error, MSE ¼ 0.856) calculated according to Mann (2004). The dashed horizontal lines are the 2 standard deviations of the period 1961–1990. The driest and the wettest Mediterranean winters for the reconstruction and the full period are denoted. precipitation reconstructions are of the order of 40 mm at 1500 and decrease to approximately 20 mm at the end of the nineteenth century. There is clear evidence of an extended dry period (with respect to the 1961–1990) at the turn of the twentieth century, followed by wet conditions with maximum in the 1960s (see also Xoplaki et al., 2004 and Chapter 3). A striking phenomenon is the negative winter rainfall trend since the 1960s (Cullen and deMenocal, 2000; Goodess and Jones, 2002; Xoplaki et al., 2004; Chapter 3), which seems to be unprecedented as inferred from this reconstructed long-term time series. This negative trend can at least partly be explained by the positive trend of the NAO (e.g. Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2004; Chapter 3 and references therein). Figure 25 clearly reveals that enhanced anomalies only appeared after 1800. This is an artefact of the statistical reconstruction approach, or a low number of proxies, rather than a real climate
Mediterranean Climate Variability over the Last Centuries: A Review
89
feature of the time period 1500–1800. Therefore, particular care is required when interpreting changes in extreme values prior to 1800. In general, the frequency and amplitude of intense anomalies from the long-term mean is steadily increasing between 1500 and 2002. The winter of 2002–03 (not shown) was distinctly wetter than the 1961–1990 average. The absolute driest (1988–89) and wettest (1962–63) winters were observed within the twentieth century. Figure 26 presents anomaly maps of the driest (1881–82, 45 mm drier than the 1961–1990 period) and the wettest (1837–38, 45 mm wetter than the 1961–1990
Figure 26: Anomalous precipitation for the wet Mediterranean winter 1837–38 (top) and the dry Mediterranean winter 1881–82 (bottom) (with respect to 1961–1990).
90 Mediterranean Climate Variability
period) winters of the reconstruction period. The wet Mediterranean winter of 1837–38 was characterized by positive precipitation anomalies stretching from southwestern Europe and northwestern Africa over Italy towards the Balkans. Drier, though not significant conditions are found from Tunisia to the Near East region and parts of Turkey. This anomaly pattern strongly resembles the correlation map between the winter NAOI and Mediterranean winter precipitation (i.e. Cullen et al., 2002; Xoplaki, 2002). Indeed, the instrumental NAOI for the 1837–38 winter (Vinther et al., 2003b) indicates a strong negative value. In the case of the dry Mediterranean winter of 1881–82 widespread negative precipitation anomalies are found over the northern Mediterranean, Iberia and northwestern Africa, whereas more precipitation was received within a band stretching from Italy, Tunisia over Greece towards Libya and the Near East. This winter was connected to a strongly positive NAOI. The wettest (driest) decade was 1961–1970 (1986–1995). The wettest (driest) multidecadal periods (30 Mediterranean winters in a row) were from 1951 to 1980 (1973–2002) with 5 mm ( 15 mm) departures from the 1961 to 1990 average. The spatial distribution of these anomalies is presented in Fig. 27. The 30 driest Mediterranean winters (1973–2002) indicate, that especially the central part, the southern Balkans and the eastern Mediterranean experienced dryness (Fig. 27, top). There are, however, areas that received more precipitation compared with 1961–1990. Concerning the 30 wettest winters over the last centuries (Fig. 27, bottom), there is no uniform distribution. Drier areas are next to regions with positive rainfall anomalies. A more sophisticated picture of changes in climate variability over the larger Mediterranean area is drawn by the wavelet spectra in Fig. 28. The method uses the Morlet wavelets (Torrence and Compo, 1998) and is designed to describe the relative importance of different timescale within different sub-periods of a time series. Mediterranean winter precipitation reveals basically two signals (top panel): At the interannual up to decadal timescales variability continuously increases, especially after 1800 with a peak during the most recent 30 years (compare Fig. 25). It is interesting to note, that Hurrell and van Loon (1997) report on the spectral peak of the winter NAO at about 6–10 years over the last decades of the twentieth century in agreement with similar spectral power found in the Mediterranean winter near-surface air temperature presented in Fig. 28 (bottom panel). In addition, there is a strong multidecadal component, which, however, is partly beyond the cone of influence (dashed line) – a sector that cannot be interpreted because of the temporal limitation of the time series. With respect to temperature (Fig. 28, bottom), variability is more equally split up
Mediterranean Climate Variability over the Last Centuries: A Review
91
Figure 27: Anomalous winter (DJF) precipitation composites. Top: Mediterranean land precipitation for the 30 driest winters in a row (1973–2002) over the last 500 years minus the 1961–1990 reference period (in mm). Bottom: As top, but for the 30 wettest winters (1951–1980 minus 1961–1990). Data are taken from Mitchell et al. (2004) and Mitchell and Jones (2005).
into different timescales. Interannual and decadal variations slightly dominate but not persist during the entire period. Interannual variability may be somewhat enhanced since 1850 (Fig. 22). The most striking feature is the multi-centennial trend since the middle of the nineteenth century.
92 Mediterranean Climate Variability
Figure 28: Morlet wavelet spectra of regional-mean winter Mediterranean precipitation and near-surface temperature from 1500 to 2002. The values denote the wavelet power. Values larger than 4 are statistically significant at the 5% level. A basic question is whether particularly dry and wet as well as cold and warm winters occurred more frequently during the twentieth century, when climate may be partly affected by human activity through emissions of GHGs (Houghton et al., 2001). The analysis of climate extremes and their changes
Mediterranean Climate Variability over the Last Centuries: A Review
93
requires a very careful procedure. There is large uncertainty in the estimate of extremes, simply because they represent infrequent events with small sample size (Palmer and Ra¨isa¨nen, 2002). In order to use the entire information from the precipitation time series, a Gamma distribution is fitted to the data. The Gamma distribution is an appropriate statistical distribution to describe precipitation amount at various timescales. The parameters of the Gamma distribution can be determined using the method of L-moments (Hosking, 1990). The Gamma distributions are fitted to the high-pass filtered time series (Fig. 29) in order to remove the effect of enhanced low-frequency variability and transient climate features. Thus, an extreme is defined as a certain departure from the decadal-mean background state. Fitting the theoretical distribution separately to running 50-year time windows between 1500 and 2002 incorporates the aspect of climate change. The top panel in Fig. 29 shows the Gamma distributions fitted to some exemplary periods. As expected, the Gamma distribution for winter precipitation is similar to a normal distribution. During the centuries, the shape of the distribution is getting broader. This implies that variability in the time windows steadily increases (Fig. 25). In addition, the mean is shifted towards a lower value in the most recent 50-year period, although the negative rainfall trend in recent decades has been removed (Fig. 25). Winters with anomalously abundant precipitation are defined by means of return values given return times of 5, 10, 20 and 50 years. Longer return periods may theoretically be derived but are subject to enhanced uncertainty, since within the reference period the distribution is based on only 50 years. The changes in the various return values over all running 50-year time windows are depicted in the middle panel of Fig. 29. It is obvious that anomalously wet winters become more frequent over the centuries (e.g. Hennessy et al., 1997; Milly et al., 2002; Christensen and Christensen, 2003). This can either be illustrated as a decrease in the return times or an increase in the return values. An adequate picture can be drawn for excessively dry winters (not shown). Note that it is still unclear to which extent this intensification of extreme winters arises from the reconstruction method or from a real climate change signal. A final issue is to estimate whether the enhanced frequency of extreme winters during recent decades in the Mediterranean region is indeed statistically significant. The linear trend is not suitable measure for extreme changes, because the latter usually do not obey a normally distributed random process (Zhang et al., 2004). Therefore, a Monte Carlo sampling is carried out in order to estimate the uncertainty range of different extreme value estimates and to compare the resulting confidence intervals with each other (Kharin and Zwiers, 2002). This approach consists of the following steps: new samples are drawn from the fitted statistical distribution. For each new sample the return values are
Figure 29: Top: Gamma (G) distributions fitted to different 50-year periods of Mediterranean winter precipitation. Middle: Time series of estimated return values between 5 and 50 years, referring to different 50-year periods between 1500 and 2002. Bottom: Statistical significance of changes in extreme annual precipitation for different return periods, testing the last 50 years (1953–2002) against all previous 50-year periods between 1500 and 1952.
Mediterranean Climate Variability over the Last Centuries: A Review
95
determined. Repeating this 1,000 times with a randomized selection of new samples leads to a distribution of estimated return values instead of one (uncertain) estimate from the given data. Typically, the return values are normally distributed over a large number of Monte Carlo samples (Park et al., 2001). Thus, the standard deviation and confidence intervals indicate the level of uncertainty of the extreme value estimate. Given two different time windows, like for instance the first and last 50 winters of the long-term time series, a change in the return values is defined to be statistically significant at a certain error level, if the corresponding confidence intervals of the return value estimates are not overlapping between both periods. For instance the 1% significance level is reached, if the 90% confidence intervals are separated from each other. In principle, this test evaluates changes in extreme values with respect to the level of uncertainty of the extreme value estimate itself. The bottom panel in Fig. 29 illustrates the statistical significance of changes in the occurrence of excessively wet winters during the period 1953–2002 compared with all previous periods until 1952. At first sight, the changes are significant at the 1% level with respect to the first 300 winters of the time series. This is not astonishing, since variability is substantially lower in the early part of the data set (Fig. 25). On the other hand, it is also evident that the last decades were not characterized by a significantly enhanced number of wet winters compared with the climate conditions between 1850 and 1952. For shorter return periods like 5 years, the changes are even less pronounced. Assuming that the low level of interannual variability is related to the reconstruction, there is no indication that excessively wet winters have significantly changed during the last centuries. The same holds for anomalously dry winters (not shown). The same method has also been applied to Mediterranean winter nearsurface temperature (Fig. 30). The basic difference is that the normal distribution is used to fit annual-mean temperatures. A widening of the distributions is also visible (top panel). Transient changes in the return values are less apparent than in the case of precipitation (middle panel; e.g. Domonkos et al., 2003). Periods with reduced and enhanced warm (and cold) winters are alternating over the centuries. A considerable shift occurred between the late eighteenth century with less pronounced warm periods and the second half of the nineteenth century with excessively warm winters. However, this shift is not part of a consistent climate change signal. This is also inferable from the bottom panel in Fig. 30: statistically significant changes in the occurrence of anomalously warm winters between the last 50 years and previous times are only found with respect to the late eighteenth century. The same holds for cold winters (not shown). This analysis of long-term time series of Mediterranean winter precipitation and temperature demonstrates that changes in extreme values cannot easily
96 Mediterranean Climate Variability
Figure 30: Same as Fig. 27 but for near-surface temperature, using the normal distribution.
Mediterranean Climate Variability over the Last Centuries: A Review
97
be detected. Although an enhancement of excessively wet and dry winters is at first sight apparent from the time series in Fig. 25, a careful estimate of the statistical significance of changes in anomalous rainfall winters reveals that the twentieth century situation does not stand out from climate conditions in previous centuries. The signal with respect to the period 1500–1800 is barely convincing, given the systematic underestimation of climate variability (and thus uncertainties) during this early reconstruction period. In terms of near-surface temperatures practically no coherent changes in the occurrence of warm and cold winters can be unmasked. Thus, there is no evidence that the behaviour of Mediterranean climate extremes at this interannual timescale is inconsistent with natural climate fluctuation during earlier centuries. Note that this does not imply that climate change signals do not exist at other timescales like for instance daily extreme events (Kostopoulou and Jones, 2005; Paeth and Hense, 2005). In addition, it is conceivable that the long-term trends during the last decades may be outstanding. Finally, this analysis is based on regional-mean time series. It is still possible that remarkable changes have occurred in many sub-regions of the Mediterranean Basin, which do not show up in the regional mean due to opposing tendencies and compensatory effects (e.g. Xoplaki, 2002; Luterbacher and Xoplaki, 2003). Another application of reconstructed temperature and precipitation over the Mediterranean can be used for applications such as the PDSI. The PDSI is an index related to the amount of water available for plants. It is normalized and calibrated for the region where it is calculated (Wells et al., 2004). PDSI is a complex combination of temperature and precipitation. It has a memory of several years, so that winter and summer PDSI are highly correlated. Winter PDSI can be considered here as the water availability at the beginning of the growing season. We have calculated the PDSI for three regions: Morocco, Italy and Greece. Figure 31 shows that the long-term (centennial) variations are quite similar between the three regions, but the decennial variations can be different. The 1660–1900 period appears to be wet (positive values of PDSI), in agreement with the flood records of Durance River (France, see Fig. 7), and we can observe a slow decrease in the water available from the beginning of the twentieth century to today. Nevertheless it is not clear for Italy, which remains wet during most of the twentieth century. The fact that the LIA was humid is certainly due to lower temperature (and then lower evaporation) during the warm season, as we have seen for Morocco that precipitation was rather low during this season. In the contrary, the low PDSI during the period before 1650 should be explained by higher temperature as Moroccan precipitation does not appear to be particularly low.
98 Mediterranean Climate Variability
N40E22.5
N37.5E15
N32.5E-5
Winter PDSI Reconstruction in different Mediterranean regions
Figure 31: Reconstruction of the winter PDSI (Palmer Drought Severity Index) at three gridpoints (Morocco: 32.5 N, 5 W; Italy: 37.5 N, 15 E; Greece: 40 N, 22.5 E) along a west–east gradient, using tree-ring series, with, in grey, the 90% confidence interval. Grey dots represent the observed series (Guiot et al., 2005).
1.5. Connection between the Large-Scale Atmospheric Circulation and Mediterranean Winter Climate over the Last Centuries This section is devoted to the connection between the winter large-scale atmospheric circulation and the Mediterranean climate over the last few centuries, the main patterns and the changing influence through time.
1.5.1. Major Atmospheric Circulation Patterns Associated with Mediterranean Winter Climate Anomalies Based on objectively reconstructed and analysed seasonal mean SLP grids for the North Atlantic European area covering the 500-year period from 1500 to 1999
Mediterranean Climate Variability over the Last Centuries: A Review
99
(Luterbacher et al., 2002b), the major circulation patterns associated with warm, cold, dry and wet Mediterranean winters have been derived according to guidelines described in detail by Jacobeit et al. (1998). At first, all temperature or precipitation winter 0.5 0.5 grids (Luterbacher et al., 2004; Pauling et al., 2005; Figs. 22, 25) 1500–2002 were averaged for the whole Mediterranean land area. Second, those winters that differ from the overall mean by more than one standard deviation have been selected as warm, cold, dry and wet seasons, respectively. For each of these samples, the corresponding seasonal mean SLP grids have been submitted to T-mode principal component analyses with varimax rotation resulting in major circulation patterns (Figs. 32–35) explaining around 90% of the SLP variances during these anomalous Mediterranean winter seasons. Due to the T-mode (columns denote SLP gridpoints, rows are the winters (years) of the Principal Component Analyses (PCA), only very few cases with reflected patterns (significantly negative time coefficients) do occur; thus, the sign of the anomalies in Figs. 32–35 is valid for nearly all cases represented by the corresponding circulation pattern (Jacobeit et al., 1998). For wet seasons, Fig. 32 depicts as most important pattern (47% of explained variance) a configuration resembling the NOAA-CPC Scandinavian pattern in its positive mode with high-pressure anomalies centred around the Baltic Sea and low-pressure anomalies covering the Mediterranean area. This blocking pattern is consistent with a negative NAOI characterizing most of the Mediterranean wet patterns. An exception is pattern 3 with its cyclonic centre above southern Great Britain reaching up to the western Mediterranean. Patterns 2–4 have less explained variance (between 10 and 18%) and a wet impact only in restricted areas of the whole Mediterranean. Mediterranean dry patterns (Fig. 33) mostly depict a positive NAO pattern except for pattern 4 that reveals an anomalous configuration with highpressure deviations from the central Mediterranean to the Icelandic region. The most important dry pattern (nearly 57% of explained variance) includes well-developed westerlies in higher mid-latitudes and anticyclonic conditions throughout the Mediterranean region. In contrast to that, pattern 3 implies an opposition between the western and the eastern Mediterranean, similar to the positive mode of a recent canonical correlation pattern, which is strongly linked to various indices of the Mediterranean Oscillation (Du¨nkeloh and Jacobeit, 2003). However, pattern 3 only accounts for 12% of the variance during the historical 500-year period. Pattern 2 which in turn is related to drier conditions in the eastern Mediterranean, has no distinct anomaly centre in the western region; thus, the Mediterranean oscillation might have been less developed during historical times than during the twentieth century. Among the Mediterranean warm patterns (Fig. 34) the positive NAO pattern known from the dry seasons recurs again as most important one (40% of
100 Mediterranean Climate Variability
Figure 32: Major four circulation patterns associated with wet Mediterranean winter seasons (DJF) in the 1500–1999 period, derived according to the procedure developed in Jacobeit et al. (1998) (using normalized T-mode principal component scores; dashed lines denote positive values).
Figure 33: Same as Fig. 32, but for dry Mediterranean winter seasons.
Mediterranean Climate Variability over the Last Centuries: A Review
101
102 Mediterranean Climate Variability
Figure 34: Same as Fig. 32, but for warm Mediterranean winter seasons.
Mediterranean Climate Variability over the Last Centuries: A Review
103
explained variance). The other patterns change the sign of the anomaly centres compared with the dry patterns thus allowing for regional warm air advection (from the SW or the SE in patterns 2 or 3, respectively) or a distinct anticyclonic regime (like in the western part of pattern 4). The Mediterranean cold patterns (Fig. 35) are similar to both the wet and the dry analyses: the most important one (41% of explained variance) largely reproduces the first wet pattern. It corresponds to a blocking configuration with easterly to northeasterly airflow towards the Mediterranean. The other patterns are known, with slight modifications, from the dry analysis, depicting different anticyclonic centres, which organize cold air advection into different parts of the Mediterranean area.
1.5.2. Mediterranean Climate Variability since the Mid-Seventeenth Century in Terms of Large-Scale Circulation Dynamics Based on objectively reconstructed monthly mean SLP grids for the North Atlantic/European area covering the period back to 1659 (Luterbacher et al., 2002b), large-scale circulation dynamics have been investigated in terms of frequency and within-type changes of major dynamical modes (Jacobeit et al., 2003). Some of these changes might have affected the Mediterranean region whose low-frequency variations in winter temperature and precipitation are reproduced in Fig. 36. Thus, the last decades of the seventeenth century became cooler and wetter in the Mediterranean area. They were marked by an increased importance of the surface Russian High pressure pattern with easterly dry and cold airflow advancing further to the west. The first half of the eighteenth century became warmer and drier in the Mediterranean region, and was dominated by large-scale westerly patterns before another period with increasing Russian High influence coincided with wetter Mediterranean conditions. Around the turn of the eighteenth to the nineteenth century westerly patterns prevailed again and led to drier and warmer winters in the Mediterranean area. Subsequently, until the mid-nineteenth century, the Russian High pressure pattern strengthened once more with easterly dominance and distinct cyclonic influence in the central Mediterranean (Jacobeit et al., 2003). This caused a period of increasing rainfall and decreasing temperatures. During the second half of the nineteenth century the Russian High retreated gradually and was replaced by different westerly patterns, which induced a Mediterranean rainfall minimum around 1900 (see also Fig. 25 above Section 1.4). Afterwards, the well-known changes of the twentieth century took place with a sustained warming and an initial rainfall increase being replaced by a sharp decline during the last four decades (Fig. 36; see also Chapter 3).
104 Mediterranean Climate Variability
Figure 35: Same as Fig. 32, but for cold Mediterranean winter seasons.
Mediterranean Climate Variability over the Last Centuries: A Review
Precipitation (mm)
105
Temperature (°C)
Figure 36: 31-year running averages of Mediterranean winter (DJF) temperature (grey line) and precipitation (black line) for the period 1500–2002.
1.5.3. Running Correlation Analysis between the Large-Scale Atmospheric Circulation and Mediterranean Winter Climate over the Last Centuries A further investigation has been made on the strength and change of the correlation between the averaged Mediterranean winter precipitation and temperature time series and atmospheric circulation for the 1764–2002 period using 30-year running correlation. This period has been reconstructed independently (only station pressure data used for the SLP reconstruction, see Luterbacher et al., 2002b; Casty et al., 2005b; Touchan et al., 2005b). Reconstructed eastern North Atlantic/European SLP fields become reliable from 1764 onwards. We calculated the first and second winter SLP Empirical Orthogonal Function (EOF) that account together for approximately 68% of total variance of European winter SLP. We estimated significance levels of the 30-year running correlations between the PCs and Mediterranean precipitation and temperature, respectively, from 1,000 Monte Carlo simulations of independent white noise processes. Thus, this analysis is an indication of potential instabilities through time and the importance of the main European atmospheric patterns on regional winter Mediterranean climate. The first EOF of winter (December–February) SLP for the period 1764–2002 explains 44% of the total winter SLP variability and reveals the well-known dipole pattern, resembling the NAO (not shown). The spatial correlation between the NAOI and Mediterranean winter temperature and precipitation for the twentieth century and interpretations is given
106 Mediterranean Climate Variability
in Cullen et al. (2002), Xoplaki (2002), Du¨nkeloh and Jacobeit (2003) and in Chapter 3. The second EOF of winter SLP (explaining around 24%) is a monopole pattern with negative anomalies centred over the British Isles (not shown). The Mediterranean area is at the southern border of this negative anomaly pattern. This pattern strongly resembles the East Atlantic/Western Russia (EATL/WRUS) pattern, which is one of the two prominent patterns that affect Eurasia during most of the year. It has been referred to as the Eurasia-2 pattern by Barnston and Livezey (1987). The spatial correlation between the EATL/WRUS and Mediterranean winter temperature and precipitation for the last decades of the twentieth century as well as its importance for the Mediterranean climate is given in Xoplaki (2002), Du¨nkeloh and Jacobeit (2003) and in Chapter 3. The second SLP Principal Component (PC) correlates significantly at the 99% level with the EATL/WRUS index over the common period 1950–2002 (not shown). Thus, this PC can be considered a ‘‘proxy’’ for the winter East Atlantic/Western Russia pattern back to 1764. Figure 37 presents the 30-year running correlation between the first and second PC of winter SLP and averaged winter Mediterranean surface air temperature 1764–2002. Except for the short period at the beginning of the nineteenth century, the correlation between the first PC (resemblance with the NAO) and the averaged Mediterranean temperature is mostly significantly negative (Fig. 37, top). It should be mentioned that there are regional differences between the NAO and Mediterranean winter temperature. For instance, the winter NAOI is significantly negatively correlated with the winter air temperature over a huge area mainly in the southeastern part, including central Algeria, Libya and Egypt, southern Italy, Greece, Turkey, Cyprus and the entire Near East countries (Cullen and deMenocal, 2000; Xoplaki, 2002). In agreement with those findings, the combined analysis of a seasonally resolved coral oxygen isotope record from the northernmost Red Sea (Felis et al., 2000) and reconstructed climate fields over the eastern North Atlantic and Europe (Luterbacher et al., 2002b) revealed the important role of the AO/NAO for eastern Mediterranean/Middle East winter climate on interannual timescales since 1750 (Rimbu et al., 2006). These findings are consistent with other evidence of a connection between the NH annular modes (AO) and eastern Mediterranean/Middle East climate variability in past centuries (Cullen and deMenocal, 2000; Felis et al., 2000; Rimbu et al., 2001; Cullen et al., 2002; Mann, 2002b; Luterbacher and Xoplaki, 2003). The combined analysis of proxy records derived from fossil corals of the northernmost Red Sea and simulations with a coupled atmosphere–ocean circulation model (ECHO-G) revealed an AO/NAO influence on the region’s interannual and mean climate during the
Mediterranean Climate Variability over the Last Centuries: A Review
107
Figure 37: 30-year running correlation between the first (top panel), and second (bottom panel) Principal Component of winter SLP and averaged winter Mediterranean surface air temperature 1764–2002. The 90 and 95% correlation significance levels have been calculated with Monte Carlo simulations. late Holocene and last interglacial period 2,900 and approximately 122,000 years ago, respectively (Felis et al., 2004). On the other hand, the areas from Iberia to Italy, the southern Balkans towards the Black Sea return non-significant correlations between the NAO and wintertime air temperature. Positive relationships are prevalent over the northwestern Mediterranean coast, central and eastern Europe with maximum values north of 45 N (Cullen and deMenocal, 2000; Xoplaki, 2002). Thus, in these running correlations there are compensatory effects between areas with negative, and positive correlations, which obviously change through time. Further, the differences might also be related to not constant transfer functions over time. Finally, the question arises to which extent climatic patterns during the twentieth century (calibration period of the statistical models) represent the entire range of climate variability (e.g. Luterbacher and Xoplaki, 2003).
108 Mediterranean Climate Variability
The bottom part of Fig. 37 clearly shows stable significantly positive correlation between the second PC of winter SLP and averaged winter Mediterranean temperature back to 1764. The EATL/WRUS pattern in its negative phase tends to bring increased anomalous southwesterly circulation connected with significant higher overall Mediterranean winter temperatures. Figure 38 presents the 30-year running correlation between the first and second PC of winter SLP and averaged winter Mediterranean precipitation 1764–2002. The upper panel of Fig. 38 indicates significantly negative correlations between the first PC of winter SLP (similar to the NAO) and averaged Mediterranean winter precipitation. The spatial correlation map between the winter NAOI and Mediterranean precipitation (Cullen and deMenocal, 2000; Xoplaki, 2002) supports these findings. Except for southeastern part of the Mediterranean the remaining areas reveal negative correlations. In terms of PC2, the running correlations are mostly positive after 1850 onwards.
Figure 38: Same as Fig. 37, but for averaged winter Mediterranean precipitation (RR).
Mediterranean Climate Variability over the Last Centuries: A Review
109
Figures 37 and 38 show that PC1 (PC2) has a robust signal on Mediterranean mean land precipitation (temperature), while the influence on Mediterranean mean land temperature (precipitation) is fluctuating and depends on the time window (not shown). It is suggested, that PC1 (PC2) impact on Mediterranean precipitation (temperature) is rather homogeneous throughout the region, while the impact on temperature (precipitation) is not so homogenous and sub-regional processes can provide different signals, which can eventually cancel.
1.6. Teleconnection Studies with Other Parts of the Northern Hemisphere Recent findings of idealized SST anomaly experiments by Hoerling et al. (2001, 2004) and Hurrell et al. (2004), indicate that SST variations have significantly controlled the North Atlantic circulation and the Mediterranean. They are related to the NAO, with the warming of the tropical Indian and western Pacific Ocean being of particular importance. A review on the influence of extratropical teleconnections and ENSO on Mediterranean climate for the recent instrumental period is provided by Chapter 2. Here we mention a few examples and the potential on possible teleconnections covering the last few centuries. A preliminary approach to identify regional patterns of teleconnections in a northern hemispheric scale, based on extreme conditions observed in a small area like Greece has been reported by Zerefos et al. (2002). They have performed superimposed epoch analysis. Winter and summer cold and warm years (25–75%-percentiles, respectively) are used as keydates. They are calculated from the Athens air temperature records on temperature NCEP reanalysis data since 1948 (Kalnay et al., 1996; Kistler et al., 2001). Their analyses included cases with dry and wet quantiles as calculated from the Athens precipitation records on mean SLP data (Trenberth and Paolino, 1980), available since 1899. The results are presented in Figs. 39 upper, lower panels–42 upper, lower panels, for winter and summer, respectively. The figures clearly show closed contours over Greece of the appropriate sign of the departure from the mean, both in near-surface air temperature and mean SLP. In the case of air temperature these anomalies are related to opposite sign of anomalies between Greece and over northwestern Europe, clearly seen in the lower quantile cases for both winter and summer. For the dry and wet cases, these opposing sign contours are more evident in summer. Moreover, in summer and in all cases examined, anomalies of the same sign as those observed over Greece and southeastern Europe appear as well over parts of South Asia. These patterns need also to be evaluated from
Figure 39: Average near-surface air temperature for cold years in Greece (lower 25%) for summer (JJA; contour interval 0.25 C), upper panel, and winter (DJF, contour interval 0.5 C), lower panel.
Figure 40: Average near-surface air temperature for warm years in Greece (upper 75%) for summer (JJA; contour interval 0.5 C), upper panel, and winter (DJF; contour interval 0.5 C), lower panel.
Figure 41: Average mean sea level pressure for dry years in Greece for summer (JJA; contour interval 0.5 hPa), upper panel, and winter (DJF; contour interval 0.5 hPa), lower panel.
Figure 42: Average mean sea level pressure for wet years in Greece summer (JJA; contour interval 0.5 hPa), upper panel, and winter (DJF; contour interval 0.5 hPa), lower panel.
112 Mediterranean Climate Variability
Table 3: Simultaneous droughts in Greece and China derived from documentary evidence. Periods of drought in Greece
Number of drought years in Henna (Northern China) within each period
305–340 551–559 741–742 1040–1046 All other years from 1 to 1909, excluding the periods 305–340, 551–559, 741–742, 1040–1046
16 drought years out of 36 (44%) 1 drought year out of 9 (11%) 0 out of 2 0 out of 7 251 drought years out of 1,855 (13.5%)
documentary records and compared to model analysis for northwestern and southeastern Europe and Asia (e.g. Hsu, 2002; Wang, 2002). The teleconnection pattern seen in Fig. 41 has been tested with independent documentary evidence mostly from monastery data (in case of Greece from Repapis et al., 1989; Xoplaki et al., 2001 and for China from the ‘‘Reconstruction of Climate Data from Ancient Chinese History’’). Preliminary results (Table 3) indicate that the teleconnection pattern is probably confirmed only for the long-lasting drought period 305–340. For comparison, the percentage of drought days in all other years from 1 to 1909 (excluding the periods 305–340, 551–559, 741–742, 1040–1046) for the same region in northern China (Henna) is 13.5%. Further analysis will also allow to investigate the relationship and its change through time between Mediterranean regions with published evidence from paleoclimate reconstructions over the last centuries to millennia in China (Wang and Zhao, 1981; Zhang and Crowley, 1989; Song, 1998, 2000; Wang et al., 2001; Qian and Zhu, 2002; Yang et al., 2002; Ge et al., 2003, 2005; Paulsen et al., 2003; Qian et al., 2003a,b; Sheppard et al., 2005). Jones (2004) and Jones et al. (2004, 2005) have recently studied a highresolution record of lake oxygen isotope change from a varved crater lake in continental central Turkey (i.e. the semi-arid Cappadocia sub-region). It records changes in summer evaporation, by comparing it with the records of Indian and African (Sahel) Monsoon rainfall at an annual resolution through the instrumental period. They have found that the relationships show periods of increased evaporation in the continental Central Anatolia region of Turkey associated with periods of increased Monsoon rainfall in India and Africa, and this relationship is also found to hold through the last 2,000 years when using comparative proxy records of both monsoon systems. According to the findings
Mediterranean Climate Variability over the Last Centuries: A Review
113
of Jones (2004) and Jones et al. (2004, 2005), the largest inferred shifts in the atmospheric circulation over this time frame occurred around 530 and 1400, and are linked to shifts between relatively warm and cold periods of the NH climate. In contrast to the stable influence of the AO/NAO over longer periods, the influence of the ENSO on southeastern Mediterranean climate is strongly non-stationary. The analysis of observational data reveals significantly positive correlations between SST anomalies in the tropical Pacific and in the eastern Mediterranean Sea/northern Red Sea during the mid-1930s to late-1960s (Rimbu et al., 2003a). The correlations jump to negative values in the 1970s. The 1970s shift in the ENSO teleconnection on southeastern Mediterranean climate, which is also documented in a coral record from the northernmost Red Sea (Felis et al., 2000), is not unique. Similar shifts occurred frequently in the last 250 years, suggesting that the ENSO impact on the region is modulated at interdecadal timescales (Rimbu et al., 2003a). There seems to be also a connection between past and present Nile flood maxima and ENSO events (Eltahir and Wang, 1999; Kondrashov et al., 2005 and references therein). The coherence between ENSO index and the Nile flood has a distinguished peak at the timescale of 4–5 years, which is close to the ENSO timescale (Eltahir and Wang, 1999). The possible physical connections are discussed in Eltahir and Wang (1999) and Kondrashov et al. (2005) as well. Further, the drought events that occurred in western Sicily during the 1565–1915 period were compared with ENSO (Piervitali and Colacino, 2001). Results show, that in periods of many drought events a reduction of ENSO events occurred and vice versa.
1.7. Mediterranean Winter Temperature and Precipitation Reconstructions in Comparison with the ECHO-G and HadCM3 Coupled Models Models can help us determine how we might have expected the climate system to change given past changes in boundary conditions and forcings, which we can compare to inferences derived from paleoclimatic data (Jones and Mann, 2004). Internal variability generated in coupled ocean–atmosphere models can be verified against the long-term variability evident in proxy-based temperature reconstructions of the past centuries. This section deals with the comparison between the 500-year winter temperature and precipitation reconstructions discussed in Section 1.4 and the ECHO-G and HadCM3 models. Two simulations have been made using the ECHO-G atmosphere–ocean GCM (Legutke and Voss, 1999). This model consists of the spectral atmospheric
114 Mediterranean Climate Variability
°
model ECHAM4 (Roeckner et al., 1996) and the ocean model HOPE-G (Wolff et al., 1997) both developed at the Max Planck Institute of Meteorology in Hamburg. The atmospheric spectral model is used with a T30 horizontal resolution (approximately 3.75 3.75 ) and 19 vertical levels. The ocean component is HOPE-G with an equivalent horizontal resolution of 2.8 2.8 and 20 vertical levels. A constant in time flux adjustment was applied to avoid climate drift. Two simulations of the ECHO-G climate model (Erik and Columbus, see Fig. 43) are produced with the same external forcing specification. Further technical details, descriptions and results with these simulations are specified in Gonza´lez-Rouco et al. (2003a,b), Zorita et al. (2003, 2004, 2005) and von Storch et al. (2004). The simulations were driven with estimations of external forcing factors (solar variability, atmospheric GHG concentrations (CO2, CH4, N2O) and radiative effects of stratospheric
±
Figure 43: Comparison between empirically reconstructed winter average Mediterranean (10 W–40 E; 30 N–47 N) land-based surface air temperature reconstructions (1500–1900), Mitchell et al. (2004), Mitchell and Jones (2005) gridded data (1901–1990) (Fig. 24) and model-based estimates over the period 1500–1990. Shown are 30-year Gaussian smoothed series with respect to the reference period 1901–1930. The simulations include full three-dimensional atmosphere–ocean general circulation models (ECHO-G (Erik, Columbus) and HadCM3) based on varying radiative forcing histories. The filtered reconstructions are presented with their 30-year filtered 95% confidence interval. See text for details.
Mediterranean Climate Variability over the Last Centuries: A Review
115
volcanic aerosols) for the period 1000 (1550) to 1990 derived from the data provided by Crowley (2000): Annual global mean concentrations of CO2 and CH4, were derived from polar ice cores (Blunier et al., 1995; Etheridge et al., 1996). The values of N2O were used as in previous scenario experiments with this model (Battle et al., 1996; Roeckner et al., 1999). Short-wave radiative forcing from solar irradiance (10Be concentrations in ice cores and 14C in tree rings and sun spots observations) and volcanic activity (from ice core acidity) are aggregated into a global annual value. Changes in ozone concentrations, land use changes, tropospheric aerosols and orbital parameters have not been considered. HadCM3 is a finite-difference model with an atmospheric horizontal resolution of 2.5 latitude 3.75 longitude and a 1.25 1.25 horizontal resolution of the oceanic component. Unlike ECHO-G, HadCM3 does not need any flux adjustment. This simulation was forced with both natural and anthropogenic forcings for the 1750–2000 period. It complements a separate run with naturalonly forcings from 1500 to 2000. The natural forcings were volcanic (for four equal-area latitude bands), orbital and solar. The anthropogenic forcings were well-mixed greenhouse gases, aerosols, tropospheric and stratospheric ozone changes, and land surface changes. See Tett et al. (2005) for more detail. The main differences in the external forcing with respect to the ECHO-G simulation are the inclusion of the effect of anthropogenic tropospheric aerosols, ozone, other trace greenhouse gases and land-use changes. The differences in sensitivity to external forcings of ECHO-G and HadCM3 are described in Brohan et al. (2005). Further differences between the two GCMs can be related to differences in their internal variability (Min et al., 2005) and to their ability of properly representing important geographical details as well as stratospheric processes which can lead to an inadequate representation of the NAO (e.g. Stendel et al., 2006). Figure 43 presents a comparison between the empirically reconstructed winter average Mediterranean (10 W–40 E; 30 N–47 N) land-based surface air temperature reconstructions 1500–1990 (Fig. 22, Section 1.4) together with associated filtered 2 standard errors and model-based estimates over the period 1500–1990. A similar comparison has been presented for the Alpine area and the European continent (Goosse et al., 2005c; Raible et al., 2005). The simulations include full three-dimensional atmosphere–ocean general circulation models (ECHO-G and HadCM3). Except for the Columbus (ECHO-G) simulation the two other simulations Erik (ECHO-G) and the HadCM3 run tend to fall well within the 2 standard error bands of the filtered reconstruction. Both ECHO-G-simulations seem to be more at variance with the reconstruction before 1900. This feature is because the period 1901–1930 is a reference for all curves and the ECHO-G
116 Mediterranean Climate Variability
integrations have larger trends than the HadCM3 starting from the midnineteenth century. The smaller trends in the HadCM3 simulation are probably due to a different climate sensitivity and the inclusion of aerosol forcing which acts to reduce GHG warming especially in summer over continental areas. Also, the response of the main North Atlantic patterns of circulation (NAO, EA, etc.) can play an important role in determining long-term regional temperature trends (e.g. Xoplaki et al., 2003) in the Mediterranean region, thus differences in the simulated temperature response in both models can also partially be related to the response of the internal dynamics in each model and simulation which should be addressed in more detail in the future. Furthermore, ensemble runs of past simulations are needed (e.g. Goosse et al., 2005c; Yoshimori et al., 2005). Prior to the twentieth century, all simulations tend to show a similar range of variability. The Columbus run presents the most extreme episodes at the end of the seventeenth century and first half of the nineteenth century. The first extreme episode in solar activity occurs during the well-known Maunder Minimum. A temperature minimum at this time can be found in both ECHO-G simulations and in the reconstruction. The second extreme episode in the Columbus simulation (around 1825) is not found in the reconstructions and the other two simulations over the Mediterranean and would be in phase with the Dalton Minimum in solar activity. Some discussion on the occurrence of these minima at hemispherical and global scales in the ECHO-G simulations and in reconstructions can be found in Gonza´lez-Rouco et al. (2003a,b), Zorita et al. (2004, 2005), and Wagner and Zorita (2005). Figure 44 presents a comparison between the empirically reconstructed winter Mediterranean (10 W–40 E; 30 N–47 N) land-based precipitation reconstructions 1500–1990 (Fig. 25) together with associated filtered 2 standard errors and model-based estimates over the period 1500–1990. It is noticeable that one of the ECHO-G simulations (Erik) falls well within the uncertainty bands of the reconstruction through the five centuries. Columbus and the HadCM3 simulations tend to exceed the lower uncertainty band. This is due to a slightly larger simulated than reconstructed variability and to the reference period used, which produces some relative shift of these simulations to lower precipitation values in the first four centuries. In general, the simulated interannual variability of winter Mediterranean precipitation seems to be larger than that of the reconstruction. The behaviour of precipitation in the larger Mediterranean area is strongly related to dynamics in the North Atlantic–European area (Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2004; Chapter 3). For specific regions, smaller scale Mediterranean cyclogenesis (see Chapter 6) plays a determinant role and other large-scale patterns as ENSO or the African Monsoon (see Chapter 2) have been named as possible sources of
Mediterranean Climate Variability over the Last Centuries: A Review
117
Figure 44: Same as Fig. 43, but for averaged Mediterranean winter land precipitation. precipitation variability. Progress in understanding the differences in simulated and reconstructed variability will come also from further assessment of the behaviour of large-scale patterns of circulation in the simulations and reconstructions (e.g. Luterbacher et al., 2002b; Casty et al., 2005b,c) as well as the limits of the model simulations in reproducing smaller scale convective precipitation. The reconstructions and the model simulations show no clear response to changes in the external forcing. This is also supported by the lack of coherence at decadal to centennial timescales between rainfall in simulations with similar versions of the model (ECHO-G) and the same external forcing. Also the correlation with the reconstructed winter NAOI is not significant, and no trend is discernible in the twentieth century. Therefore, the internal variability seems to be larger than the potential signal caused by variations in the external forcing (Yoshimori et al., 2005). At lower frequencies, some response of the NAOI to GHGs is evidenced by model simulations (Zorita et al., 2005).
1.7.1. Past Climate Variability and its Relations to Volcanism, Solar Activity and GHG Concentrations: Reconstructions and Models Recent studies (Lionello et al., 2005) have compared in detail model results from the ECHO-G simulations of the past centuries with the spatial temperature reconstruction for the European and Mediterranean region (Luterbacher et al.,
118 Mediterranean Climate Variability
2004), focusing on the relative role of the external forcing and of the NAO dynamics. Only winter is considered in the comparison. The model simulation and the reconstruction present important differences, which do not allow reaching certain conclusions on the role of the Solar Volcanic Radiative Forcing (SVRF) on the Mediterranean climate. According to the model dynamics, the SVRF only slightly correlates with winter European temperature at multidecadal timescales and the two forced simulations show a disagreement over northern Europe and northern Africa (not shown). In fact, during this season, the effects of the internal dynamics are superimposed to that of the SVRF forcing. Consequently, the average winter temperature difference between the two simulations is well correlated (0.59) with the difference between their winter NAOIs. The reconstruction shows a different behaviour. Correlation between reconstructed temperatures and SVRF is significant only at the multidecadal timescales over northern Europe, while over southern Europe values are negative. Consequently, the correlation between simulations and reconstruction is low over most of southern Europe (not shown, Lionello et al., 2005). An overestimation of the climate sensitivity to SVRF by the ECHO-G model is possible: its results are likely to be very inaccurate also because of the poor representation of the regional details of the Mediterranean region. Another possibility is that the SVRF might contain some error, or its spatial variability, which is not explicitly described in the experiments, might have important effects in southern Europe. Finally, inaccuracies in the reconstruction cannot be ruled out. Identifying the relationships between the main regional climatic patterns and the radiative forcing due to solar activity, volcanic eruptions and GHG concentration is a key point to assess the model capability to predict future scenarios at the regional scale. A low correlation between radiative forcing (RF) and regional climate and the lack of interactions between RF and the internal climate variability would limit the possibility to reconstruct and to predict regional climate, especially in the Mediterranean region. In general, if present results are confirmed, a considerable part of the Mediterranean temperature changes is associated with the internal, unpredictable climate dynamics, and simulations of future climate scenarios could be skilful only at the multidecadal timescales. Additional modeling studies of the response to solar and volcanic forcing have been done with the NASA Goddard Institute for Space Studies (GISS) models. These have been used to examine the spatial patterns of the equilibrium response to solar irradiance changes and both the short- and long-term response to volcanism in a historical context (Shindell et al., 2001a, 2003, 2004; Schmidt et al., 2004). Many of these studies were performed with a relatively coarse resolution (8 by 10 ) version of the GCM with 23 layers and a mixed-layer
Mediterranean Climate Variability over the Last Centuries: A Review
119
ocean. The model emphasized the representation of stratospheric dynamics, and included a detailed treatment of the forcings such that volcanic aerosols were specified with latitudinal and vertical time-varying distributions and solar irradiance variations were spectrally resolved. The latter plays a significant role in solar forcing since most of the variability is at short wavelengths that is primarily absorbed in the stratosphere. The model also included the chemical response of stratospheric ozone to solar variability. Focusing on the Maunder Minimum period, the model produced the generally cooler conditions of the late seventeenth century in comparison with the late eighteenth century (Shindell et al., 2003). Solar forcing provided the dominant contribution, largely by introducing a prolonged negative wintertime NAO phase (Luterbacher et al., 1999, 2002a), consistent with earlier studies and with proxy reconstructions (Shindell et al., 2001a). Volcanic forcing contributed to a general cooling during this period, but with little spatial structure at multidecadal scales. Stratospheric ozone’s response to solar variability played an important role in the outcome in those simulations, amplifying the large-scale dynamic changes by around a factor of two. The results matched large-scale proxy network reconstructions in many areas of the NH, including the cooling in northern Europe during the late seventeenth century (e.g. Luterbacher et al., 2004; Xoplaki et al., 2005). As noted previously, the decade with the coldest Mediterranean winter temperatures in the proxy-based reconstruction was 1680–1689 (Section 1.4, Fig. 22). However, the model failed to reproduce the opposite polarity anomaly centered over the eastern Mediterranean. Newer simulations with a more sophisticated version of the GISS GCM called modelE have been carried out for the case of volcanic eruptions. The simulations, run at 4 5 resolution and again with 23 vertical layers, reveal that the model is able to capture the spatial pattern of the short-term response to volcanic eruptions much better than the earlier model (Shindell et al., 2004). This pattern is almost identical to the longer term solar response pattern, and the newer model captures both the northern European (Luterbacher et al., 2004; Xoplaki et al., 2005) and opposite polarity eastern Mediterranean responses (Shindell et al., 2004). Since this pattern is also thought to occur largely via the same processes as the longer term solar forcing, this raises the possibility that models may soon be able to reproduce historic Mediterranean climate variability much better than they have in the past.
1.8. Conclusions The larger Mediterranean area offers a remarkably high quality and quantity of long instrumental series, a wide range of documentary data (Section 1.2.1)
120 Mediterranean Climate Variability
as well as high and low spatio-temporally resolved natural proxies (tree rings, speleothems, corals and boreholes; Section 1.2.2). The recognition of correlated chemical and isotopic signals in new marine archives (e.g. non-tropical coral, the vermetid reefs and the deep-sea corals) provide an important new approach to help unravel climatic variability in the Mediterranean basin for the last few centuries (Section 1.2.2). This multi-proxy climate information make the larger Mediterranean area very suitable for climate reconstructions at different timescales, as well as the analysis of changes in climate extremes and socio-economic impacts prior to the instrumental period. However, there are still a lot of archives with documentary proxy evidence unexplored, as has been shown for the Iberian Peninsula, Italy, France, the Balkans, Greece, the southeastern Mediterranean area, possibly the northern African countries as well as ship logbooks. Further, the coverage of northern Africa with respect to natural, annually resolved proxy climate data is sparse. Exceptions are tree-ring records from Morocco (work in Algeria and Tunisia just started) and coral records from the northernmost Red Sea (Egypt, Israel, Jordan). To the best of our knowledge no annually resolved proxy climate data exists so far from the Mediterranean coastal region of Libya and Egypt. More tree-ring work in these regions as well as non-tropical corals and vermetid reefs could help to fill this gap. Further, there is potential in the eastern Mediterranean, mainly in Turkey, for new speleothem data which may provide information on the precipitation conditions over the last centuries to millennia. Approximately one-third of the country is underlain by carbonate rocks and caves are frequent. Those caves are geographically well distributed, making a north–south and west–east transect across Turkey possible (D. Fleitmann, personal communication). The data generated by the various studies presented and discussed in this review contribute to the regional, national and international scientific and resource management communities. Internationally, these data will fill critical gaps in multiple global climate databases valued by programmes such as PAGES, CLIVAR, NOAA-OGP and NASA-EOS. For instance, tree-ring evidence from the southeastern Mediterranean is especially useful for investigations of interannual to multi-century-scale climate variability, ranging from climate system oscillations to recurrent regional drought episodes. Regionally, it is important to provide a decadal to multi-century perspective on climate variability to local managers of land and water resources. Given the significance of water in the region and its strategic importance, this could have considerable impact on international water resource policy, management and political agreements between countries in the region. The Mediterranean is a water-deficit region and in parts there is a history of conflict over land and natural resources. This information will aid in anticipating and, hopefully lessening the likelihood of conflict over scarce water resources.
Mediterranean Climate Variability over the Last Centuries: A Review
121
We addressed the question on the importance and limitations of documentary and natural proxies for Mediterranean precipitation reconstructions at seasonal timescale. We found that different proxy types have their specific response region, which suggests to use region-specific multi-proxy sets in seasonal climate reconstructions and confirm recent findings devoted to near-surface air temperature (Pauling et al., 2003). Documentary precipitation and tree-ring data are those proxies, which are the most important both for boreal winter and summer precipitation covering large areas around the Mediterranean. However, other proxies such as corals, speleothems and ice cores are relevant for smaller restricted areas. In order to verify the preliminary conclusions more systematic testing of a larger dataset of proxies is needed, e.g. additional data around the Mediterranean area with seasonal resolution considering additional data around the Mediterranean areas resolving seasonal resolution. It should also be taken into account that not only the proxy type determines the results but also its initial number, location, quality and availability over time. Numerous seasonally resolved documentary proxy data and information gathered from natural proxies discussed in Sections 1.2 and 1.3 have been used to derive winter Mediterranean temperature and precipitation fields and averaged time series back to 1500 (Section 1.4). It turned out that rather moderate seasonal to multidecadal precipitation and temperature variability was prevalent over the last few centuries. Several cold relapses and warm intervals as well as dry and wet periods on decadal timescales, on which shorter period quasi-oscillatory behaviour was superimposed have been found. In the context of the last half millennium, however, the last winter decades of the twentieth/twenty-first century were the warmest and driest. The analysis of anomalously wet and warm winters in Section 1.4 has revealed that in the regional-mean time series of Mediterranean winter precipitation and temperature no statistically significant changes with respect to the frequency and intensity of extreme winters have occurred since 1500. The relationship between large-scale atmospheric circulation patterns and Mediterranean winter climate anomalies during the last 500 years (Section 1.5) may be generalized in the following way: warm and dry winters linked with positive NAO modes (first patterns in Fig. 32); cold and wet winters are connected with Scandinavian blocking (first patterns in Fig. 33). Cold and dry winters could be related to different anticyclonic regimes (patterns 2–4 in Fig. 34), whereas warm and wet winters are connected with different cyclonic regimes (patterns 2–4 in Fig. 35). Running correlation analyses between the leading patterns derived from the large-scale atmospheric circulation and the regional averaged Mediterranean temperature and precipitation reveals, that the NAO (East Atlantic/ Western Russia) has a robust signal on Mediterranean mean land
122 Mediterranean Climate Variability
precipitation (temperature), whereas the influence on Mediterranean mean land temperature (precipitation) is fluctuating and depends on the time window. It is suggested, that PC1 (PC2) of large-scale SLP impact on Mediterranean precipitation (temperature) is rather homogeneous throughout the region, while the impact on temperature (precipitation) is not so homogenous and sub-regional processes can provide different signal which can eventually cancel (the signal is non-significant for the last few centuries). Section 1.6 showed that there are indications for possible teleconnections between Mediterranean and NH climate in past centuries. Finally, the comparison among the reconstructions of Mediterranean winter temperature and land precipitation with the ECHO-G and HadCM3 simulations (Section 1.7) shows that the range of variability reproduced by the climate models for the period 1500–1990 is only slightly larger than that of the reconstructions (Section 1.4). In the case of temperature, the HadCM3 simulation trends are comparable with those in the empirical reconstructions (Section 1.4) and slightly smaller than those in the ECHO-G simulations. This is a reasonable feature, since the latter do not include aerosol forcings and land use changes. As for the case of Mediterranean precipitation, no trends are reconstructed nor simulated (Fig. 44). The variability of Mediterranean precipitation is also slightly larger for the simulations compared to the reconstructions. Both assessments reveal the need for a more thorough study that takes into consideration the behaviour of the atmospheric circulation in climate reconstructions and model simulations.
1.9. Outlook Despite the fact that there are many proxy data available from the larger Mediterranean area, the uncertainties of the climate reconstructions increase back in time. In order to improve reconstruction skill in time and space and expand climate estimates further back in time, one main aim is therefore to enlarge the spatio-temporal coverage of high-resolution, high-quality, accurately dated, natural and documentary proxy evidence from all countries along the Mediterranean Sea, as there are many sources which have not yet explored and regions with scarce information (i.e. North African coastal regions) and those sensitive to climate change. Archives of the Islamic world, yet unexplored, are believed to provide documentary evidence on past weather and climate. Attention should also be paid to ship logbooks of which many thousands have now been located, as a major and reliable data source points to a more comprehensive review of climatic variation in the region than has hitherto been
Mediterranean Climate Variability over the Last Centuries: A Review
123
the case. Another interesting source of data is related to agricultural production, which may be inferred from the taxes paid by farmers and recorded in municipal and ecclesiastic accounts books. New high-resolution marine archives and application of isotopic and geochemical proxies will be of much relevance for past SST, salinity, near-surface air temperature and rainfall reconstructions. Such new archives will provide weekly to annual resolution records for the Mediterranean Sea, using rather low-cost sampling and analytical techniques. More research and tree-ring sampling in the entire Mediterranean region are required as tree-ring information is probably the most important proxy for large areas around the Mediterranean. For instance, even though many chronologies were developed from the southeastern Mediterranean, this is not enough to faithfully represent an area that expands almost 2,000 km from east to west and 1,500 km from north to south. This remains one of the largest mid-latitude semi-arid regions without an adequate network of climate-sensitive tree-ring chronologies. Recently, R. Touchan started a large-scale systematic sampling in Morocco, Tunisia and Algeria. This project (supported by NSF-ESH) will establish a multi-century network of North African climate records based on tree rings, by extending and enhancing the existing tree-ring dataset geographically and temporally. This network will then be used to study interannual to centuryscale climate fluctuations in the region, and their links to large-scale patterns of climate variability. Up to date 15 chronologies from Morocco and Tunisia were developed. The length of the chronologies range from 113 (Dahallia, Tunisia) to 1082 (Col Du Zad, Morocco) years. The incorporation of the multi-proxy data with a high spatio-temporal coverage, together with sophisticated reconstruction methodologies (bearing in mind the extreme character of some of the data, linear and non-linear approaches) will provide a broader picture of past Mediterranean climate variability covering the last few centuries, not only averaged over the entire area but for specific sub-regions including the Mediterranean Sea. An important issue for future investigation on past temperature and precipitation extreme is the application of the analysis technique described in Section 1.4 to shorter timescales and individual sub-regions of the Mediterranean basin in order to infer whether significant changes of opposite sign may have happened at different sites. Further analyses, however, have to take into consideration sub-areas as current conditions clearly indicate different impacts of major teleconnection patterns within different Mediterranean regions. The combination of highly resolved climate reconstructions and model climate simulations offer extended scientific understanding on the climate response to external forcing (e.g. the direct radiative and the poorly investigated dynamical response to tropical eruptions over the Mediterranean). Another approach to draw a better picture of past climate variability over the Mediterranean area covering the last millennia is to combine high- and
124 Mediterranean Climate Variability
low-resolution climate proxies using sophisticated reconstruction methods as demonstrated for the NH by Moberg et al. (2005). Future work involving climate reconstructions and model simulations can benefit our understanding of Mediterranean climate variability at several levels. Model simulations can be used as a pseudoreality in which gridpoint variables can be degenerated with noise and treated as pseudoproxies to replicate reconstruction methods within the simulated world (Mann and Rutherford, 2002; Zorita and Gonza´lez-Rouco, 2002; Gonza´lez-Rouco et al., 2003a, 2006; Zorita et al., 2003; von Storch et al., 2004; Mann et al., 2005). The simulation of relevant large-scale climate regimes for Mediterranean variables can be comparatively studied with reconstructions of atmospheric circulation. This should provide knowledge about the potential response of relevant circulation regimes to external forcing and aid in understanding differences among simulations and reconstructions (Casty et al., 2005b,c). Further, the range of simulated variability for precipitation and temperature in simulations of the last millennium in comparison with multi-proxy reconstructions can help estimate uncertainties in scenario simulations of future climate change. In addition, the assessment of the deficiencies of model simulations in reproducing temperature and precipitation at smaller scales, in particular, that related to Mediterranean cyclogenesis, can help improve model parameterizations. Understanding of variability at smaller spatial scales and a better simulation of involved processes can be achieved with regional climate modeling and statistical downscaling. Model studies investigating the role of processes, such as the stratospheric ozone chemical response to solar variability or the ocean circulation response to solar or volcanic perturbations, can help to determine how regional patterns are setup and why model results differ at regional scales. Last, but not least, the assessment of the changes in the dynamics associated to extreme climate episodes through the last millennium registered both in climate simulations and in empirical reconstructions will help better understand the mechanisms involved in extremes and related impacts on society and economy. As shown in this review, the Mediterranean is a region with evident singularities in pluviometric patterns. It produces relatively frequently climatic hazards like floods and droughts. Due to demographic and touristic developments, people are increasingly exposed to ‘‘climatic hazards’’. Thus, climatic research must be focused on both, modeling and reconstruction of extreme meteorological and climatic events. For reconstruction of rarely occurring extreme events, research on historical documentary sources and fluvial/ lacustrine sediments for instance is needed to identify, analyse and reconstruct them and put them in a long-term context. Only with this knowledge urban and regional planning can produce long-term strategies for climatic hazards prevention.
Mediterranean Climate Variability over the Last Centuries: A Review
125
Acknowledgements Ju¨rg Luterbacher, Elena Xoplaki, Carlo Casty and Erich Fischer are supported by the Swiss National Science Foundation (NCCR). Elena Xoplaki, Joel Guiot, Simon Tett, Andreas Pauling, Eduardo Zorita were financially supported through the European Environment and Sustainable Development programme, project SOAP (EVK2-CT-2002-00160). Manola Brunet, Elena Xoplaki and Jucundus Jacobeit were financially supported through the European Environment and Sustainable Development programme, project EMULATE (EVK2-CT-2002-00161). Jucundus Jacobeit and Ju¨rg Luterbacher thank the SFN-Floodrisk/DFG-Extreme1500 for research on the SLP reconstruction and synoptic analysis. Stefan Bro¨nnimann was funded by the Swiss National Science Foundation (Contract Number PP002-102731). This Rutishauser is supported by the Swiss National Science Foundation (Contract Number 205321-105691/1). Mariano Barriendos and Fidel Gonzalez-Rouco acknowledge support from the ‘‘Research Programme Ramon y Cajal, Ministery of Education and Science, Spain’’. Mariano Barriendos would like to thank also the ‘‘Research Project REN2002-04584-C04-03/CLI, Ministry of Education and Science, Spain’’. The studies of the Italian climate performed by CNR ISAC (Dario Camuffo and collaborators) were funded by the European Commission, DGXII (Environment and Climate Programme) and CORILA, Venice, Italy. Simon Tett was funded by the UK Government Met. Research (GMR) contract. Computer time for the HadCM3 simulations was funded by the UK Department for Environment, Food and Rural Affairs under the Climate Prediction Program Contract PECD 7/12/3. Thomas Felis was supported by Deutsche Forschungsgemeinschaft through DFG-Research Center for Ocean Margins at Bremen University, contribution No. RCOM0309. Ramzi Touchan was supported by the US National Science Foundation, Earth System History (Grant No. 0075956). Sergio Silenzi and Paolo Montagna were supported by the Italian Institute for Marine Research (ICRAM) under the Paleoclimate Reconstruction program; they also wish to thank Ermenegildo Zegna Foundation for financial support and Claudio Mazzoli and Marco Taviani for comments and suggestions. Michele Brunetti, Teresa Nanni and Maurizio Maugeri acknowledge support by CLIMAGRI (Italian Ministry for agriculture and forests), ALP-IMP (EU-FP5), and U.S.–Italy bilateral Agreement on Cooperation in Climate Change Research and Technology (Italian Ministry for the environment). Jose Carlos Gonzalez Hidalgo acknowledges support by ‘‘Ministerio de Ciencia y Tecnologı´ a Research Project REN2002-01023-CLI’’. Norel Rimbu was supported by the AWI through the MARCOPOLI(MAR2) program. We wish to thank the reviewers for their constructive comments that improved the quality of the chapter.
126 Mediterranean Climate Variability
References Adkins, J. F., Griffin, S., Kashgarian, M., Cheng, H., Druffel, E. R. M., Boyle, E. A., Edwards, R. L., & Shen, C.-C. (2002). Radiocarbon dating of deep-sea corals. Radiocarbon, 44, 567. Adkins, J. F., Henderson, G. M., Wang, S.-L., O’Shea, S., & Mokadem, F. (2004). Growth rate of the deep-sea scleractinia Desmophyllum cristagalli and Enallopsammia rostrata. Earth Planet. Sc. Lett., 227, 481. Akkemik, U¨. (2000). Dendroclimatology of umbrella pine (Pinus pinea L.) in Istanbul, Turkey. Tree-Ring Bull., 56, 17. Akkemik, U¨., & Aras, A. (2005). Reconstruction (1689-1994 AD) of April-August precipitation in the southern part of central Turkey. Int. J. Climatol., 25, 537. Akkemik, U¨., Dag˘deviren, N., & Aras, A. (2005). A preliminary reconstruction (A.D. 1635-2000) of spring precipitation using oak tree-rings in the western Black Sea region of Turkey. Int. J. Biomet., 49, 297, doi: 10.1007/s00484-004-0249-8. Alcoforado, M. J., Nunes, M. F., & Garcia, J. C. (1997). Climate and society in Portugal, before the setting of an institutional meteorological network. Publ. AIC, 10, 75, (in French). Alcoforado, M. J., Nunes, M. F., & Garcia, R. (1999). The perception of the relation between climate and public health in Lisbon, on the 19th century based on Franzini (1779–1861) writings, Revista Portuguesa de Sau´de Pu´blica 17, Lisboa, 31 (in Portuguese). Alcoforado, M. J., Nunes, M. F., Garcia, J. C., & Taborda, J. P. (2000). Temperature and precipitation reconstruction in southern Portugal during the late Maunder Minimum (1675–1715). Holocene, 10, 333. Amades, J. (1984). Histo`ries i llegendes dels carrers de Barcelona, Edicions 62, Barcelona, 2 vols. Angot, A. (1885). Etude sur les Vendanges en France, Annales du Bureau Central M0 et0 eorologique de France, publie´es par E. Mascart. Ann0 ee 1883. I.0 Etude des orages en France et M0 emoires divers. Paris, B.29–B.120 1623. Angulo, R. J., Giannini, P. C. F., Suguio, K., & Pessenda, L. C. R. (1999). Relative sea-level changes in the last 5500 years in Southern Brazil (Laguna-Imbituba region, Santa Catarina State) based on vermetid 14C ages. Mar. Geol., 159, 23. Antonioli, F., Chemello, R., Improta, S., & Riggio, S. (1999). Dendropoma lower intertidal reef formations and their palaeoclimatological significance, NW Sicily. Mar. Geol., 161, 155. Antonioli, F., Cremona, G., Immordino, F., Puglisi, C., Romagnoli, C., Silenzi, S., Valpreda, E., & Verrubbi, V. (2002). New data on the holocenic sea level rise in NW Sicily (Central Mediterranean Sea). Global Planet. Change, 716, 121. Antonioli, F., Silenzi, S., & Frisia, S. (2001). Tyrrhenian holocene palaeoclimate trends from spelean serpulids. Quaternary Sci. Rev., 20, 1661. Antonioli, F., Silenzi, S., Gabellini, M., & Mucedda, M. (2003). High resolution climate trend over the last 1000 years from a stalagmite in Sardinia (Italy). Quaternaria Nova, 7, 1. Arago, F. (1858). L’Etat thermome´trique du globe terrestre. Oeuvres completes, Paris. Araus, J. L., Febrero, A., Buxo, R., Camalich, M. D., Martin, D., Molina, F., Rodriguezariza, M. O., & Romagosa, I. (1997). Changes in carbon isotope discrimination in grain cereals from different regions of the Western Mediterranean Basin
Mediterranean Climate Variability over the Last Centuries: A Review
127
during the past seven millennia. Palaeoenvironmental evidence of a differential change in aridity during the late Holocene. Global Change Biol., 3, 107. Ayalon, A., Bar-Matthews, M., & Kaufman, V. (1999). Petrography, trace elements (Ba, Sr, Mg and U) and isotope geochemistry of strontium and uranium in speleothems as paleoclimate proxies: Soreq Cave Israel. Holocene, 9, 715. Baker, A., Ito, E., Smart, P. L., & McEwan, R. (1997). Elevated 13C in speleothem and implications for palaeovegetation studies. Chem. Geol., 136, 263. Baldini, J. U. L., McDermott, F., & Fairchild, I. J. (2002). Structure of the 8200-year cold event revealed by a speleothem trace element record. Science, 296, 2203. Bannister, B. (1970). Dendrochronology in the Near East: current research and future potentialities. Proc. of the seventh International Congress of Anthropological and Ethnological Sciences, 5, 336. Bard, E., Delaygue, G., Rostek, F., Antonioli, F., Silenzi, S., & Schrag, D. P. (2002). Hydrological conditions over the Western Mediterranean Basin during the deposition of the cold Sapropel 6 (ca. 175 kyr BP). Earth Planet. Sc. Lett., 202, 481. Bar-Matthews, M., Ayalon, A., & Kaufman, A. (1997). Late Quaternary palaeoclimate in the Eastern Mediterranean region from stable isotope analysis of speleothems at Soreq Cave, Israel. Quaternary Res., 47, 155. Bar-Matthews, M., Ayalon, A., & Kaufman, A. (2000). Timing and hydrological conditions of sapropel events in the Eastern Mediterranean, as evident from speleothems, Soreq Cave, Israel. Chem. Geol., 169, 145. Bar-Matthews, M., & Ayalon, A. (2004). Speleothems as paleoclimate indicators, a case study from Soreq Cave located in the eastern mediterranean region, Israel. In: R. W. Battarbee et al. (Eds), Past climate variability through Europe and Africa (Springer, Academic Publishers, Dordrecht, Germany, 638 pages, chapter 18). Barnston, A. G., & Livezey, R. E. (1987). Classification, seasonality and persistence of low frequency atmospheric circulation patterns. Mon. Weather Rev., 115, 1825. Barriendos, M. (1997). Climatic variations in the Iberian peninsula during the late Maunder minimum (AD 1675-1715): an analysis of data from rogation ceremonies. Holocene, 7, 105. Barriendos, M. (2005). Climate and culture in Spain, religious responses to extreme climatic events in the Hispanic Kingdoms (16th–19th centuries). In: W. Behringer, H. Lehmann, & C. Pfister (Eds), Cultural Consequences of the Little Ice Age (Kulturelle Konsequenzen der Kleinen Eiszeit) (Vandenhoech & Ruprecht, Go¨ttingen, pp. 31–86). Barriendos, M., & Llasat, M. C. (2003). The case of the ‘‘Malda´’’ anomaly in the western Mediterranean Basin (AD 1760–1800): an example of a strong climatic variability. Climatic Change, 61, 191. Barriendos, M., & 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. Barriendos, M., & Rodrigo, F. S. (2005). Seasonal rainfall variability in the iberian peninsula from the 16th century: preliminary results from historical documentary sources. Geophys. Res. Abstracts, 7. Barriendos, M., Martin-Vide, J., Pena, J. C., & Rodriguez, R. (2002). Daily meteorological observations in Cadiz – San Fernando. Analysis of the documentary sources and the instrumental data content (1786–1996). Climatic Change, 53, 151.
128 Mediterranean Climate Variability
Bartholy, J., Pongracz, R., & Molnar, Z. (2004). Classification and analysis of past climate information based on historical documentary sources for the Carpathian basin. Int. J. Climatol., 24, 1759. Battarbee, R.W., Gasse, F., Stickley, C.E., Past climate variability through Europe and Africa, Series: Developments in Paleoenvironmental Research, Vol. 6 (28 chapters, 638 pages), 2004, Springer, Germany. Battle, M., Bender, M., Sowers, T., Tans, P. P., Butler, J. H., Elkins, J. W., Ellis, J. T., Conway, T., Zhang, N., Lang, P., & Clarke, A. D. (1996). Atmospheric gas concentrations over the past century measured in air from firn and the South Pole. Nature, 383, 231. Bauer, E., Claussen, M., Brovkin, V., & Huenerbein, A. (2003). Assessing climate forcings of the Earth system for the past millennium. Geophys. Res. Lett., 30, 1276, doi: 10.1029/2002GL016639. Beltrami, H. (2002). Earth’s long-term memory. Science, 297, 206. Beltrami, H., & Bourlon, E. (2004). Ground warming patterns in the Northern Hemisphere during the last five centuries. Earth and Planet. Sci. Lett., 227, 169. Beltrami, H., & Mareschal, J. C. (1995). Resolution of ground temperature histories inverted from borehole temperature data. Global Plan. Change, 11, 57. Beltrami, H., Chapman, D. S., Archambault, S., & Bergeron, Y. (1995). Reconstruction of high resolution ground temperature histories combining dendrochronological and geothermal data. Earth. Planet. Sci. Lett., 136, 437. Beltrami, H., Ferguson, G., & Harris, R. N. (2005). Long-term tracking of climate change by underground temperatures. Geophys. Res. Lett., 32, L19707, doi: 10.1029/ 2005GL023714. Benito, G., Dı´ ez-Herrero, A., & Ferna´ndez de Villalta, M. (2003a). Magnitude and frequency of flooding in the Tagus Basin (Central Spain) over the last millennium. Climatic Change, 58, 171. Benito, G., Sopena, A., Sanchez-Moya, Y., Machado, M. J., Perez-Gonzalez, A., & Ferna´ndez de Villalta, M. (2003b). Paleoflood record of the Tagus river (Central Spain) during the late pleistocene and holocene. Quarternary Sci. Rev., 22, 1737. Bivona-Bernardi, A. (1832). Caratteri dei vermeti desunti da cinque specie che abitano nel mare di Palermo. Effemeridi scientifiche e letterarie per la Sicilia 1, 2, 3. Blunier, T., Chappellaz, J. A., Schwander, J., Stauffer, B., & Raynaud, D. (1995). Variations in atmospheric methane concentration during the Holocene epoch. Nature, 374, 46. Bradley, R. S., & Jones, P. D. (1993). ‘‘Little Ice Age’’ summer temperature variations: their nature and relevance to recent global warming trends. Holocene, 3, 367. Bra´zdil, R., & Dobrovolny, P. (2000). Chronology of strong winds in the Czech Lands during the 16th–19th centuries, Instytut Geograficzny UJ. Prace Geograficzne, 107, 65. Bra´zdil, R., & Dobrovolny, P. (2001). History of strong winds in the Czech Lands: causes, fluctuations, impacts. Geographia Polonia, 74, 11. Bra´zdil, R., Glaser, C. R., Pfister, C., Dobrovolny, P., Antoine, J.-M., Barriendos, M., Camuffo, D., Deutsch, M., Enzi, S., Guidoboni, E., Kotyza, O., & Sanchez Rodrigo, F. (1999). Flood events of selected European rivers in the sixteenth century. Climatic Change, 43, 239. Bra´zdil, R., Dobrovolny, P., Elleder, L., Kakos, V., Kotyza, O., Mackova, J., Stekl, J., Tolasz, R., & Valasek, H. (2003). History of weather and climate in the Czech Lands VII. Floods. Manuscript, Brno, Czech Republic, unpaginated.
Mediterranean Climate Variability over the Last Centuries: A Review
129
Bra´zdil, R., Dobrovolny, P., Stekl, J., Kotyza, O., Valasek, H., & Jez, J. (2004). History of weather and climate in the Czech Lands VI: Strong winds. Masaryk University of Bern, Czech Republic. 377 pages. Bra´zdil, R., Pfister, C., Wanner, H., von Storch, H., & Luterbacher, J. (2005). Historical climatology in Europe – the state of the art. Climatic Change, 70, 363. Briffa, K. R., & Osborn, T. J. (2002). Blowing hot and cold. Science, 292, 662. Briffa, K. R., Jones, P. D., Schweingruber, F. H., & Osborn, T. J. (1998). Influence of volcanic eruptions on Northern Hemisphere summer temperatures over 600 years. Nature, 393, 450. Briffa, K. R., Osborn, T. J., Schweingruber, F. H., Jones, P. D., Shiyatov, S. G., & Vaganov, E. A. (2001). Low-frequency temperature variations from a northern treering density network. J. Geophys. Res., 106, 2929. Briffa, K. R., Osborn, T. J., Schweingruber, F. H., Jones, P. D., Shiyatov, S. G., & Vaganov, E. A. (2002). Tree-ring width and density data around the Northern Hemisphere: part 2, spatio-temporal variability and associated climate patterns. Holocene, 12, 759. Briffa, K. R., Osborn, T. J., & Schweingruber, F. H. (2004). Large-scale temperature inferences from tree-rings: a review. Global Planet. Change, 40, 11. Brohan, P., Tett, S., & Zorita, E. (2005). Sensitivity to external forcings of HadCM3 and ECHO-G, submitted. Bro¨nnimann, S. (2003). Picturing climate change. Clim. Res., 22, 87. Bro¨nnimann, S., & Luterbacher, J. (2004). Reconstructing Northern Hemisphere upper-level fields during World War II. Clim. Dyn., 22, 499–510. Bru¨ckner, E., (1890). Klimaschwankungen seit 1700 nebst Bemerkungen u¨ber die Klimaschwankungen der Diluvialzeit. Wien (Vienna). Brunetti, M., Buffoni, L., Lo Vecchio, G., Maugeri, M., & Nanni, T., (2001). In: M. Brunetti, L. Buffoni, G. Lo Vecchio, M. Maugeri, & T. Nanni (Eds), Tre Secoli di Meteorologia a Bologna (CUSL, Milan, Italy, pp. 95). Budillon, F., Esposito, E., Iorio, M., Pelosi, N., Porfido, S., & Violante, C. (2005). The geological record of storm events over the last 1000 years in the Salerno Bay (Southern Tyrrhenian Sea); new proxy evidences. Adv. in Geosciences, 2, 123. Bu¨ntgen, U., Esper, J., Frank, D. C., Nicolussi, K., & Schmidhalter, M. (2005a). A 1052year tree-ring proxy for Alpine summer temperatures. Clim. Dyn., 25, 141, doi: 10.1007/ s00382-005-0028-1. Bu¨ntgen, U., Esper, J., Frank, D. C., & Nievergelt, D., (2005b). Alpine summer temperature variations, AD 755–2004, J. Climate, revised. Burns, S. J., Fleitmann, D., Mudelsee, M., Neff, U., Matter, A., & Mangini, A. (2002). A 780-year annually resolved record of Indian Ocean monsoon precipitation from a speleothem from South Oman. J. Geophys. Res., 107, 4434, doi: 10.1029/ 2001JD001281. Camuffo, D. (1984). Analysis of the series of precipitation at Padova, Italy. Climatic Change, 6, 57. Camuffo, D. (1987). Freezing of the Venetian Lagoon since the 9th century AD in comparison to the climate of Western Europe and England. Climatic Change, 10, 43. Camuffo, D. (1993). Analysis of the sea surges at Venice from AD 782 to 1990. Theor. Appl. Climatol., 47, 1. Camuffo, D., & Enzi, S. (1991). Locust invasions and climatic factors from the middle Ages to 1800. Theor. and Appl. Climatol., 43, 43.
130 Mediterranean Climate Variability
Camuffo, D., & Enzi, S. (1992). Reconstructing the climate of Northern Italy from archive sources. In: R. S. Bradley, and P. D. Jones (Eds), Climate since 1500 AD, (Routledge, London). Camuffo, D., & Enzi, S., (1994a). The climate of Italy from 1675 to 1715. In: B. Frenzel (Ed), Climatic Trends and Anomalies in Europe 1675–1715, Paleoclimate Research (Fischer Verlag, Stuttgart, Special Issue 8, pp. 243–254). Camuffo, D., & Enzi, S. (1994b). Chronology of ‘dry fogs’ in Italy, 1374-1891. Theor. Appl. Climatol., 50, 31. Camuffo, D., & Enzi, S. (1995a). Climatic features during the Spo¨rer and Maunder Minima. In: B. Frenzel (Ed), Solar Output and Climate During the Holocene (Fischer Verlag, Stuttgart, Special Issue 16, pp. 105–125). Camuffo, D., & Enzi, S. (1995b). The analysis of two bi-millenary 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 Verlag, Stuttgart, Vol. 41, pp. 433–450). Camuffo, D., & Enzi, S. (1995c). Impact of clouds of volcanic aerosols in Italy in the past centuries. Nat. Hazards., 11, 135. Camuffo, D., & Enzi, S. (1996). The analysis of two bi-millennary 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 Verlag, Stuttgart, Vol. 41, pp. 433–450). Camuffo, D., Secco, C., Brimblecombe, P., & Martin-Vide, J. (2000a). Sea storms in the Adriatic Sea and the Western Mediterranean during the last millennium. Climatic Change, 46, 209. Camuffo, D., Cocheo, C., & Enzi, S. (2000b). Seasonality of instability phenomena (hailstorms and thunderstorms) in Padova, Northern Italy, from archive and instrumental sources from AD 1300 to 1989. Holocene, 10, 651. Camuffo, D. (2002a). History of the long series of the air temperature in Padova (1725-today). Climatic Change, 53, 7. Camuffo, D. (2002b). Calibration and instrumental errors in early measurements of air temperature. Climatic Change, 53, 297. Camuffo, D. (2002c). Errors in early temperature series arising from changes in style of measuring time, sampling schedule and number of observations. Climatic Change, 53, 331. Camuffo, D., & Sturaro, G. (2003). Sixty-cm submersion of Venice discovered thanks to Canaletto’s paintings. Climatic Change, 58, 333. Camuffo, D., & Sturaro, G. (2004). Use of proxy-documentary and instrumental data to assess the risk factors leading to sea flooding in Venice. Global Planet. Change, 40, 93. Camuffo, D., G. Sturaro, G., & Benito, G. (2003). An opposite flood pattern teleconnection between the Tagus Iberian Peninsula) and Tiber (Italy) rivers during the last 1000 years. In: V. R. Thorndycraft, G. Benito, M. Barriendos, and M. C. Llasat (Eds), Paleofloods and Climatic Variability: Applications in Flood Risk Management, (CSIC-CCM, Madrid, Spain, pp. 295–300). Camuffo, D., Pagan, E., & Sturaro, G. (2005). The extraction of Venetian sea-level change from paintings by Canaletto and Bellotto. In: C. A. Fletcher, and T. Spencer (Eds), Flooding and Environmental Challenges for Venice: State of Knowledge (Cambridge University Press, Cambridge, pp. 129–140).
Mediterranean Climate Variability over the Last Centuries: A Review
131
Casty, C., Luterbacher, J., Wanner, H., Esper, J., & Bo¨hm, R. (2005a). Temperature and precipitation variability in the European Alps since 1500, Int. J. Climatol., 25, 1855. Casty, C., Handorf, D., & Sempf, M. (2005b). Combined winter climate regimes over the North Atlantic/European sector 1766–2000. Geophys. Res. Lett., 32, L13801. Casty, C., Handorf, D., Raible, C.C., Gonza´lez-Rouco, J. F., Weisheimer, A., Xoplaki, E., Luterbacher, J., Dethloff, K., & Wanner, H. (2005c). Recurrent climate winter regimes in reconstructed and modelled 500 hPa geopotential height fields over the North Atlantic/European sector 1659–1990, Clim. Dynam., 24, 809, doi: 10.1007/s00382-004-0496-8. Chbouki, N., Stockton, C. W., & Myers, D. (1995). Spatio-temporal patterns of drought in Morocco. Int. J. Climatol., 15, 187. Cheng, H., Adkins, J. F., Edwards, R. L., & Boyle, E. A. (2000). 230Th dating of deep-sea corals. Geochim. Cosmochim. Ac., 64, 2401. Christensen, J. H., & Christensen, O. B. (2003). Severe summertime flooding in Europe. Nature, 421, 805. Chuine, I., Yiou, P., Viovy, N., Seguin, B., Daux, V., & Le Roy Ladurie, E. (2004). Historical phenology: Grape ripening as a past climate indicator. Nature, 432, 289. CLIWOC Team (2003). CLIWOC multilingual meteorological dictionary, An English-Spanish-Dutch-French dictionary of wind terms used by mariners from 1750–1850, KNMI publication 205, HISKLIM 5. Cocheo, C., & Camuffo, D. (2002). Corrections of systematic errors and data homogenisation in the Padova series (1725–today). Climatic Change, 53, 77. Cook, E. R., Esper, J., & D’Arrigo, R. (2004). Extra-tropical Northern Hemisphere temperature variability over the past 1000 years. Quaternary Sci. Rev., 23, 2063. Correia, A., & Safanda, J. (2001). Ground surface temperature history at a single site in southern Portugal reconstructed from borehole temperatures. Global Planet. Change, 19, 155. Creus-Novau, J., Ferna´ndez, A., & Manrique, E. (1997). Dendrocronologı´ a y clima del u´ltimo milenio en Espan˜a. Aspectos metodolo´gicos y avance de resultados: El paisaje mediterra´neo a trave´s del espacio y del tiempo. Implicaciones en la desertificacio´n, Eds. Iban˜ez, Valero Garce´s and Machado, Logron˜o, Geoforma. Creus-Novau, J., Ge´nova, M., Ferna´ndez, A., & Pe´rez, A. (1992). New dendrochronologies for Spanish Mediterranean zone. Lunqua, 34, 76. Creus-Novau, J., Zozaya, B., & Fernandez Cancio, A. (1995). Reconstrucciones climaticas en Galicia durante las ultimas centurias: estudio dendrocronologico. Xunta de Galicia, Santiago de Compostela, Spain. Crowley, T. J. (2000). Causes of climate change over the past 1000 years. Science, 289, 270. Crowley, T. J., & Lowery, T. S. (2000). How warm was the Medieval Warm period? A comment on ‘man-made versus natural climate change’. Ambio, 39, 51. Cullen, H. M., & deMenocal, P. B. (2000). North Atlantic influence on Tigris-Euphrates streamflow. Int. J. Climatol., 20, 853. Cullen, H. M., Kaplan, A., Arkin, P. A., & deMenocal, P. B. (2002). Impact of the North Atlantic Oscillation on Middle Eastern climate and streamflow. Climatic Change, 55, 315. D’Arrigo, R. D., & Cullen, H. M. (2001). A 350-year (AD 1628–1980) reconstruction of Turkish precipitation. Dendrochronologia, 19, 167.
132 Mediterranean Climate Variability
Daveau, S. (1997). Weather types in Coimbra, Portugal (Dez 1663–Set 1665) from the letters of Father Anto´nio Vieira, Finisterra- Revista Portuguesa de Geografia, XXXII, 64, 109 (in Portuguese). Davis, B. A. S., Brewer, S., Stevenson, A. C., & Guiot, J. (2003). The temperature of Europe during the Holocene reconstructed from pollen data. Quaternary Sci. Rev., 22, 1701. Delongeville, R., Laborel, J., Pirazzoli, P., Sanlaville, P., Arnold, M., Bernier, P., Evin, J., & Montaggioni, L. (1993). Les variations recentes de la ligne de rivage sur le littoral Syrien. Quaternaire, 4, 45. De Putter, T., Loutre, M. F., & Wansard, G. (1998). Decadal periodicities of Nile river historical discharge (A.D. 622–1470) and climatic implications. Geophys. Res. Lett., 25, 3193. Desmarchelier, J. M., Goede, A., Ayliffe, L. K., McCulloch, M. T., & Moriarty, K. (2000). Stable isotope record and its palaeoenvironmental interpretation for a late middle Pleistocene speleothem from Victoria Fossil Cave, Naracoorte, South Australia. Quaternary Sci. Rev., 19, 763. Desprat, S., Goni, M. F. S., & Loutre, M. F. (2003). Revealing climatic variability of the last three millennia in northwestern Iberia using pollen influx data. Earth Planet. Sc. Lett., 213, 63. Diodato, N. (2006). A 425-year Precipitation History from Documentary Weather Anomalies and Climate Records at Palermo – Astronomical Observatory (Sicily, Italy), PAGES-News 14/1, in press. Domonkos, P., Kysely´, J., Piotrowicz, K., Petrovic, P., & Likso, T. (2003). Variability of extreme temperature events in south-central Europe during the 20th century and its relationship with large-scale circulation. Int. J. Climatol., 23, 987. Dufour, M. L., (1870), Proble`me de la variation du climat. Bulletin de la Socie´te´ Vaudoise des Sciences Naturelles. 10, 359. Du¨nkeloh, A., & Jacobeit, J. (2003). Circulation dynamics of Mediterranean precipitation variability 1948-98. Int. J. Climatol., 23, 1843. Eltahir, E. A. B., & Wang, G. (1999). Nilometers, El Nino, and climate variability. Geophys. Res. Lett., 26, 489. Enzi, S., & Camuffo, D. (1995). Documentary sources of sea surges in Venice from AD 787 to 1867. Nat. Hazards, 12, 225. Esper, J., Cook, E. R., & Schweingruber, F. H. (2002). Low-frequency signals in long tree-ring chronologies for reconstructing past temperature variability. Science, 295, 2250. Esper, J., Frank, D. C., & Wilson, R. J. S. (2004). Climate reconstructions-low frequency ambition and high frequency ratification. EOS, 85, 113, 120. Esper, J., Frank, D. C., Wilson, R. J. S., & Briffa, K. R. (2005a). Effect of scaling and regression on reconstructed temperature amplitude for the past millennium. Geophys. Res. Lett., 32, L07711, doi: 10.1029/2004GL021236. Esper, J., Wilson, R. J. S., Frank, D. C., Moberg, A., Wanner, H., & Luterbacher, J. (2005b): Climate: past ranges and future changes. Quaternary Sci. Rev., 24, 2164. Etheridge, D., Steele, L. P., Langenfelds, R., Francey, R., Barnola, J. M., & Morgan V. (1996). Natural and anthropogenic changes in atmospheric CO2 over the last 1000 years from air in Antarctic ice and firn. J. Geophys. Res., 101, 4115–4128. Fairchild, I. J., Baker, A., Borsato, A., Frisia, S., Hin-ton, R. W., McDermott, F., & Tooth, A. F. (2001). Annual to sub-annual resolution of multiple trace-element trends in speleothems. J. Geol. Soc., London, 158, 831.
Mediterranean Climate Variability over the Last Centuries: A Review
133
Felis, T., Lohmann, G., Kuhnert, H., Lorenz, S. J., Scholz, D., Pa¨tzold, J., Al-Rousan, S. A., & Al-Moghrabi, S. M. (2004). Increased seasonality in Middle East temperatures during the last interglacial period. Nature, 429, 164. Felis, T., Pa¨tzold, J., Loya, Y., Fine, M., Nawar, A. H., & Wefer, G. (2000). A coral oxygen isotope record from the northern Red Sea documenting NAO, ENSO, and North Pacific teleconnections on Middle East climate variability since the year 1750. Paleoceanography, 15, 679. Felis, T., & Pa¨tzold, J. (2004). Climate reconstructions from annually banded corals. In: M. Shiyomi, H. Kawahata, H. Koizumi, A. Tsuda, and Y. Awaya (Eds), Global Environmental Change in the Ocean and on Land (Terrapub, Tokyo, pp. 205–277). Finch, A. A., Shaw, P. A., Weedon, G. P., & Holmgren, K. (2001). Trace element variation in speleothem aragonite: potential for palaeoenvironmental reconstruction. Earth Planet. Sc. Lett., 186, 255. Fischer, T. (1879). Studien u¨ber das Klima der Mittelmeerla¨nder, Erga¨nzungsheft zu Petermann’s Mittheilungen 58, 1. Fischer, E. et al. (2006) European climate response to major tropical volcanic eruptions over the last five centuries, submitted. Fleitmann, D., Burns, S. J., Mudelsee, M., Neff, U., Kramers, J., Mangini, A., & Matter, A. (2003). Holocene forcing of the Indian monsoon recorded in a stalagmite from Southern Oman. Science, 300, 1737. Fleitmann, D., Burns, S. J., Neff, U., Mudelsee, M., Mangini, A., & Matter, A. (2004). Palaeoclimatic interpretation of high-resolution oxygen isotope profiles derived from annually laminated speleothems from Southern Oman. Quaternary Sci. Rev., 23, 935. Focke, J.W. (1977). The effect of a potentially reef-building vermetid community on an eroding limestone coast, Curacao, Netherland Anthilles. Proc. of the 3rd International Coral Reef Symposium, Miami, vol. 1, 239. Fraedrich, K., & Bantzer, C. (1991). A note on fluctuations of the Nile river flood levels (715–1470). Theor. Appl. Climatol., 44, 167. Frank, N., Paterne, M., Ayliffe, L., van Weering, T., Henriet, J.-P., & Blamart, D. (2004). Eastern North Atlantic deep-ocean corals: tracing upper intermediate water 14C during the Holocene. Earth Planet. Sc. Lett., 219, 297. Franklin, B. (1784). Meteorological imaginations and conjectures. Memoirs of the Literary and Philosophical Society of Manchester, 2, 373. Frisia, S., Borsato, A., Preto, N., & McDermott, F. (2003). Late Holocene annual growth in three Alpine stalagmites records the influence of solar activity and the North Atlantic Oscillation on winter climate. Earth Planet. Sc. Lett., 216, 411. Frisia, S., Borsato, A., Fairchild, I.J., & Susini, J. (2005) Variations in atmospheric sulphate recorded in stalagmites by synchrotron micro-XRF and XANES analyses, Earth Planet. Sc. Lett., doi:10.1016/j.epsl.2005.03.026. Gallego, D., Garcia-Herrera, R., Ribera, P., & Jones, P. D. (2005). Seasonal mean pressure reconstruction for the North Atlantic (1750–1850) based on early marine data. Climate of the Past Discussions, 1, 57, SRef-ID: 1814-9359/cpd/2005-1-57. Galli, M., Guadalupi, M., Nanni, T., Ruggiero, L., & Salerno, A. (1994). Proxy winter temperatures from Ravenna Pinus pinea forest, 1653–1985. In: B. Frenzel (Ed), Climatic trends and anomalies in Europe, (Gustav Fischer Verlag, Stuttgart, Jena, New York). Garcı´ a-Herrera, R., Garcia, R. R., Hernandez, E., Prieto, M. R., Gimeno, L., & Diaz, H. (2003a). The use of Spanish historical archives to reconstruct climate variability. Bull. Americ. Met. Soc., 84, 1025.
134 Mediterranean Climate Variability
Garcı´ a-Herrera, R., Macı´ as, A., Gallego, D., Herna´ndez, E., Gimeno, L., & Ribera, P. (2003b). Reconstruction of the precipitation in the Canary Islands for the period 1595–1836. B. Am. Meteorol. Soc., 84, 1037. Garcı´ a-Herrera, R., Ko¨nnen, G. P., Wheeler, D., Prieto, M. R., Jones, P. D., & Koek, F. B. (2005a). CLIWOC: A climatological database for the world’s oceans 1750–1854, Climatic Change, 73, 1. Garcı´ a-Herrera, R., Wilkinson, C., Koek, F. B., Prieto, M. R., Calvo, N. & Herna´ndez, E. (2005b). Description and general background to ships’ logbooks as a source of climatic data, Climatic Change, 73, 13. Garnier, E. (2004). Quatre sie`cles de tempeˆtes en foreˆt. Les phe´nome`nes e´oliens majeurs en France entre les XVIe et XIXe sie`cles. Rendez-Vous Techniques, ONF, Paris, pp. 15–20. Garnier, E. (2005a). Gifts and refusals of the nature: the Cerdagne (Pyrenees) in 18th century. In: J. L. Tissier (Ed), Paysages, territoires et ame´nagements dans le sud de la France (e´ditions du CTHS, Paris, pp. 60–73). Garnier, E. (2005b). Salvatgines et Socie´te´s montagnardes catalanes. Une guerre d’extermination pluri-se´culaire (XVe-XIXe sie`cles). Actes du colloque International RESOPYR, Presses Universitaires de Perpignan, pp. 327–343. Gassner, G., & Christiansen-Weniger, F., (1942). Dendroklimatologische Untersuchungen u¨ber die Jahresringentwicklung der Kiefern in Anatolien. Nova Acta Leopoldina: Abhandlung der Kaiserlich Leopoldinisch-Carolinisch deutschen Akademie der Naturforscher NF, Band 12, Nr 80. Ge, Q. S., Zheng, J. Y., Fang, X., Man, Z., Zhang, X., Zhang, P., & Wang, W.-C. (2003). Winter half-year temperature reconstruction for the middle and lower reaches of the Yellow River and Yangtze River, China, during the past 2000 years. Holocene, 13, 933. Ge, Q. S., Zheng, J. Y., Hao, Z. X., Zhang, P. Y., & Wang, W. C. (2005). Reconstruction of historical climate in China, high-resolution precipitation data from Qing Dynasty archives. B. Am. Meteorol. Soc., 86, 671. Gerber, S., Joos, F., Bru¨gger, P., Stocker, T. F., Mann, M. E., Sitch, S., & Scholze, M. (2003). Constraining temperature variations over the last millennium by comparing simulated and observed atmospheric CO2. Clim. Dynam., 20, 281. Glaser, R. (2001). Klimageschichte Mitteleuropas. 1000 Jahr Wetter, Klima, Katastrophen. Primus Verlag, Wissenschaftliche Buchgesellschaft, Darmstadt, 227 pages. Glaser, R., & Stangl, H. (2005). Climate and floods in Germany since AD 1000: data, methods, results and consequences, Surv. Geophys., 25, 485. Glaser, R., Bra´zdil, R., Pfister, C., Dobrovolny, P., Barriendos, M., Bokwa, A., Camuffo, D., Kotyza, O., Limanowka, D., Ra´cz, L., & Rodrigo, F. S. (1999). Seasonal temperature and precipitation fluctuations in selected parts of Europe during the sixteenth century. Climatic Change, 43, 169. Glueck, M. F., & Stockton, C. W. (2001). Reconstruction of the North Atlantic Oscillation, 1429-1983. Int. J. Climatol., 21, 1453. Goldstein, S. J., Lea, D. W., Chakraborty, S., Kashgarian, M., & Murrell, M. T. (2001). Uranium-series and radiocarbon geochronology of deep-sea corals: implications for Southern Ocean ventilation rates and the oceanic carbon cycle. Earth Planet. Sc. Lett., 193, 167. Gonza´lez-Rouco, J. F., von Storch, H., & Zorita, E. (2003a). Deep soil temperature as a proxy for surface air-temperature in a coupled model simulation of the last thousand years. Geophys. Res. Lett., 30, 2116, doi: 10.1029/2003GL018264.
Mediterranean Climate Variability over the Last Centuries: A Review
135
Gonza´lez-Rouco, J. F., Zorita, E., Cubasch, U., von Storch, H., Fischer-Bruns, I., Valero, F., Montavez, J. P., Schlese, U., & Legutke, S. (2003b). Simulating the climate since 1000 AD with the AOGCM ECHO-G, ESA SP-535, 329. Gonza´lez-Rouco, J. F., Beltrami, H., Zorita, E., & von Storch, H. (2006). Simulation and inversion of borehole temperature profiles in simulated climates: spatial distribution and surface coupling. Geophys. Res. Lett., 33, L01703, doi:10:1029/ 2005GL024693. Goodess, C. M., & Jones, P. D. (2002). Links between circulation and changes in the characteristics of Iberian rainfall. Int. J. Climatol., 22, 1593. Goosse, H., Renssen, H., Timmermann, A., & Bradley, R. S. (2005a). Natural and forced climate variability during the last millennium: a model-data comparison using ensemble simulations. Quaternary. Sci. Rev., 24, 1345. Goosse, H., Crowley, T. J., Zorita, E., Ammann, C. M., Renssen, H., & Driesschaert, E. (2005b). Modelling the climate of the last millennium: What causes the differences between simulations? Geophys. Res. Lett., 32, doi:10.1029/2005GL022368. Goosse, H., Renssen, H., Timmermann, A., Bradley, R. S., & Mann, M. E. (2005c). Using paleoclimate proxy-data to select and optimal realisation in an ensemble of simulations of the climate of the past millennium, Clim. Dynam., in press. Grove, A. T. (2001). The ‘‘Little Ice Age’’ and its geomorphological consequences in Mediterranean Europe. Climatic Change, 48, 121. Grove, J. M., & Conterio, A. (1994). Climate in the Eastern and Central Mediterranean: Climatic Trends and Anomalies in Europe. In: B. Frenzel (Ed), Paleoclimate Research, (Fischer Verlag, Stuttgart, pp. 1675–1715), Special Issue 8. Grove, J. M., & Conterio, A. (1995). The climate of Crete in the sixteenth and seventeenth centuries. Climatic Change, 30, 223. Guilbert, X. (1994). Les crues de la Durance depuis le XIVe`me sie`cle. Fre´quence, pe´riodicite´ et interpre´tation pale´o-climatique. Me´moire de maıˆtrise de Ge´ographie. Universite´ d’Aix-Marseille I, Aix-en-Provence, 350 p. Guiot, J. (1992). The combination of historical documents and biological data in the reconstruction of climate variations in space and time. In: B. Frenzel, C. Pfister, & B. Gla¨ser (Eds), Climate and Man, Pala¨oklimaforschung/Paleoclimate Research 7 (Special Issue EFS Project, 2, pp. 93. Guiot, J., Alleaume, S., Nicault, A., & Brewer, S. (2005). The Mediterranean droughts during the last 650 years: reconstruction from tree-rings and climate model simulation. Geophys. Res. Abstracts, 7, 02471, European Geosciences Union, Vienna, 24–29 April 2005. Guiot, J., Nicault, A., Rathgeber, C., Edouard, J. L., Guibal, F., Pichard, G., & Till, C. (2005). Last-millennium summer-temperature variations in Western Europe based on proxy data. Holocene, 15, 489, doi:10.1191/0959683605hl819rp. Gu¨nther, S. (1886). Lehrbuch der Geophysik, Vol. 2. Stuttgart. Hadfield, M. G., Kay, E. A., Gillette, M. U., & Lloyd, M. C. (1972). The Vermetidae (Mollusca Gastropoda) of the Hawaiian Islands. Mar. Biol., 12, 81. Hansen, J. E., Ruedy, R., Sato, M., Imhoff, M., Lawrence, W., Easterling, D., Peterson, T., & Karl, T. (2001). A closer look at United States and global surface temperature change. J. Geophys. Res., 106, 23947. Harris, R. N., & Chapman, D. S. (2001). Mid-latitude (30–60 N) climatic warming inferred by combining borehole temperatures with surface air temperatures. Geophys. Res. Lett., 28, 747.
136 Mediterranean Climate Variability
Hartmann, A., & Rath, V. (2005). Uncertainties and shortcomings of ground surface temperature histories derived from inversion of temperature logs. J. Geophys. Engineer., 2, 299, doi: 10.1080/1742-2132/2/4/S02. Hennessy, K. J., Gregory, J. M., & Mitchell, J. F. B. (1997). Changes in daily precipitation under enhanced greenhouse conditions. Clim. Dynam., 13, 667. Hoerling, M. P., Hurrell, J. W., & Xu, T. (2001). Tropical origin for recent North Atlantic climate change. Science, 292, 90. Hoerling, M. P., Hurrell, J. W., Xu, T., Bates, G. T., & Phillips, A. (2004). Twentieth century North Atlantic climate change. Part II: understanding the effect of Indian Ocean warming. Clim. Dynam., 23, 391. Hosking, J. R. M. (1990). L-moments: analysis and estimation of distributions using linear combinations of order statistics. J. R. Statist. Soc. B., 52, 105. Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van der Linden, P. J., Dai, X., Maskell, K., & Johnson, C. A. (Eds) (2001). Climate Change 2001. The Scientific Basis (Cambridge University Press, Cambridge, UK). Hsu, D., (2002). Workshop on historical climate reconstruction over East Asia, Session: 2, Beijing, China, October 14–16, 2002. Huang, S. (2004). Merging information from different resources for new insights into climate change in the past and future. Geophys. Res. Lett., 31, 13205, doi: 10.1029/ 2004GL019781. Huang, S., & Pollack, H. N. (1998). Global borehole temperature data-base for climate reconstruction, http://www.geo.lsa.umich.edu/climate, Data Contrib. Ser. 1998-044, IGBP PAGES/World Data Cent.-A for Paleoclimatol., NOAA/NGDC Paleolimatol. Program, Boulder, Colorado. Huang, S., Pollack, H. N., & Shen, P. Y. (2000). Temperature trends over the past five centuries reconstructed from borehole temperatures. Nature, 403, 756. Huang, Y., Fairchild, I. J., Borsato, A., Frisia, S., Cassidy, N. J., McDermott, F., & Hawkesworth, C. J. (2001). Seasonal variations in Sr, Mg and P in modern speleothems (Grotta di Ernesto, Italy). Chem. Geol., 175, 429. Hughes, M. K., Kuniholm, P. I., Garfin, G. M., Latini, C., & Eischeid, J. (2001). Aegean tree-ring signature years explained. Tree-Ring Res., 57, 67. Hurrell, J. W., & van Loon, H. (1997). Decadal variations in climate associated with the North Atlantic Oscillation. Climatic Change, 36, 301. Hurrell, J. W., Hoerling, M. P., Phillips, A. S., & Xu, T. (2004). Twentieth century North Atlantic climate change. Part I: assessing determinism. Clim. Dynam., 23, 371. Ideler, J. L. (1832), Ueber die angeblichen Vera¨nderungen des Klima, Annalen der Erd-, Vo¨lker- und Staatenkunde (Berghaus’ Annalen) 5, 417. Jacobeit, J., Beck, C., & Philipp, A. (1998). Annual to decadal variability in climate in Europe-Objectives and results of the German contribution to the European climate research project ADVICE, Wu¨rzburger Geographische Manuskripte 43. Jacobeit, J., Wanner, H., Luterbacher, J., Beck, C., Philipp, A., & Sturm, K. (2003). Atmospheric circulation variability in the North-Atlantic-European area since the mid-seventeenth century. Clim. Dynam., 20, 341, doi: 10.1007/s00382-002-0278-0. Jalut, G., Amat, A. E., Mora, S. R. I., Fontugne, M., Mook, R., Bonnet, L., & Gauquelin, T. (1997). Holocene climatic changes in the Western Mediterranean: installation of the Mediterranean climate. Comptes rendus de l’Academie des Sciences Serie Ii Fascicule a- Sciences De La Terre Et Des Planetes, 325, 327.
Mediterranean Climate Variability over the Last Centuries: A Review
137
Jalut, G., Amat, A. E., Bonnet, L., Gauquelin, T., & Fontugne, M. (2000). Holocene climatic changes in the Western Mediterranean, from south-east France to south-east Spain. Palaeogeogr. Palaeocl., 160, 255. Jones, B., & Hunter, I. (1995). Vermetid buildups from Grand Cayman British West Indies. J. Coastal Res., 4, 973. Jones, M. D. (2004). High-resolution records of climatic change from lacustrine stable isotopes through the last two millennia in western Turkey. Ph.D. thesis, University of Plymouth. Jones, M. D., Roberts, N., Leng, M. J., Tu¨rkes, M., & Moyeed, R. (2004). Eastern Mediterranean-Indian-African summer climate connections through the past 2000 years, Geophys. Res. Abstracts 6, Sref-ID: 1607-7962/gra/EGU04-A-00418. Jones, M. D., Leng, M. J., Roberts, C. N., Tu¨rkes° , M., & Moyeed, R. (2005). A coupled calibration and modelling approach to the understanding of dry-land lake oxygen isotope records. J. Paleolimnol., 34, 391. Jones, P. D., & Mann, M. E. (2004). Climate over past millennia. Rev. Geophys., 42, 2002, doi: 10.1029/2003RG000143. Jones, P. D., & Salmon, M. I. (2005). Preliminary reconstructions of the North Atlantic Oscillation and the Southern Oscillation index from wind strength measures taken during the CLIWOC period. Climatic Change, 73, 131. Jones, P. D., Jonsson, T., & Wheeler, D. (1998). Extension to the North Atlantic Oscillation using early instrumental pressure observations from Gibraltar and southwest Iceland. Int. J. Climatol., 17, 1433. Jones, P. D., Osborn, T. J., & Briffa, K. R. (2001). The evolution of climate over the last millennium. Science, 292, 662. Kalnay, E. et al. (1996). The NCEP/NCAR 40-year reanalysis project. B. Am. Meteorol. Soc., 77, 437. Katz, R. W., & Brown, B. G. (1992). Extreme events in changing climate: variability is more important than averages. Climatic Change, 21, 289. Kharin, V. V., & Zwiers, F. W. (2002). Changes in extremes in an ensemble of transient climate simulations with a coupled atmosphere-ocean GCM. J. Climate, 13, 3760. Kemp, M., & Laborel, J. (1968). Formations de vermets et d’alguescalcaires sur les cotes du Brasil, Recueil des Travaux de la Station Marine d’Endoume, 43, 9. Kistler, R. et al. (2001). The NCEP-NCAR 50-year reanalysis: monthly means CD-ROM and documentation. B. Am. Meteorol. Soc., 82, 247. Kohl, T. (1997). Paleoclimatic temperature signals – can they be washed out? Tectonophysics, 291, 225. Kolodny, Y., Bar-Matthews, M., Ayalon, A., & McKeegan, K. D. (2003). A high spatial resolution 18O profile of a speleothem using an ionmicroprobe. Chem. Geol., 197, 21. Kondrashov, D., Feliks, V., & Ghil, M. (2005). Oscillatory modes of extended Nile River records (A. D. 622-1922). Geophys. Res. Lett., 32, 10702, doi: 10.1029/ 2004GL022156. Ko¨nnen, G. P., & Koek, F. B. (2005). Description of the CLIWOC database, Climatic Change, 73, 117. Kostopoulou, E. & Jones, P. D. (2005). Assessment of climate extremes in the Eastern Mediterranean. Meteorol. Atmos. Phys., 89, 69. Kuniholm, P. I. (1990). The archaelogical record: evidence and non-evidence for climate change: the Earth’s climate and variability of the sun over recent millennia, (Eds) S. J. Runcorn & J.-C., Phil. Trans. R. Soc. Lond., A, 645.
138 Mediterranean Climate Variability
Kuniholm, P. I. (1994). Long tree-ring chronologies for the Eastern Mediterranean, Proc. of the 29th International Symposium on Archeometry. Kuniholm, P. I., & Striker, C. L. (1987). Dendrochronological investications in the Aegean and neighboring regions, 1983–1986. J. Field Archaeol., 14, 385. Laborel, J., & Delibrias, G. (1976). Niveaux marins recents a vermeti-daedu littoral ouest Africain. Bulletin de Liaison-Association Senegalaise pour l’Etude du Quaternaire Africain 47, 97. Lachenbruch, A. H., & Marshall, B. V. (1986). Changing climate: geothermal evidence from permafrost in the Alaskan Arctic. Science, 234, 689. Legutke, S., & Voss, R. (1999). ECHO-G, The Hamburg atmosphere-ocean coupled circulation model, DKRZ-Report No 18. Le Roy Ladurie, E. (1966). Paysans du Languedoc. SEVPEN, Paris. Le Roy Ladurie, E. (1983). Histoire du climat depuis l’An Mil, (Champs Flammarion, Paris). Le Roy Ladurie, E. (2004). Histoire humaine et compare´e du climate. Canicules et glaciers XIIIe`me-XVIIIe`me sie`cles. Editions Fayard, France, 740 pages. Le Roy Ladurie, E. (2005). Canicule, fraıˆ cheurs, vendages (France, 15–19th sie`cles). C. R. Biologies, 328, 213. Le Roy Ladurie, E., & Baulant, M. (1980). Grape harvests from the fiveteenth through the nineteenth centuries. J. Interdisciplinary Hist., X, 4, 839. Le Roy Ladurie, E., & Baulant, M. (1981). Grape harvests from the fifteenth through the nineteenth centuries. In: R. I. Rotberg, and T. K. Rabb (Eds), Studies in interdisciplinary history (Princeton University Press, New Jersey). Lionello, P., De Zolt, S., Luterbacher, J., & Zorita, E. (2005). Is the winter European climate of the last 500 years conditioned by the variability of the solar radiation and volcanism? GKSS Report, 2005/10, pp. 41. Llasat, D. C. M. (2004). Recent and historical Mediterranean floods (Spain, France, Italy), consequences, learnings and projects. Houille Blanche-Revue Internationale de L’eau 6, 37. Llasat, D. C. M., Rigo, T., & Barriendos, M. (2003). The Montserrat-2000 flash-flood event: a comparison with the floods that have occurred in the northeastern Iberian Peninsula since the 14th century. Int. J. Climatol., 23, 453. Luque, J. A., & Julia`, R. (2002). Lake sediment response to land-use and climate change during the last 1000 years in the oligotrophic Lake Sanabria (Northwest of Iberian Peninsula). Sediment. Geol., 148, 343. Luterbacher, J., & Xoplaki, E. (2003). 500-year winter temperature and precipitation variability over the Mediterranean area and its connection to the large-scale atmospheric circulation. In: H.-J. Bolle (Ed), Mediterranean Climate-Variability and Trends, (Springer Verlag, Berlin, Heidelberg, New York). Luterbacher, J., Schmutz, C., Gyalistras, D., Xoplaki, E., & Wanner, H. (1999). Reconstruction of monthly NAO and EU indices back to AD 1675. Geophys. Res. Lett., 26, 2745. Luterbacher, J., Xoplaki, E., Dietrich, D., Jones, P. D., Davies, T. D., Portis, D., Gonza´lez-Rouco, J. F., von Storch, H., Gyalistras, D., Casty, C., & Wanner, H. (2002a). Extending North Atlantic Oscillation reconstructions back to 1500. Atmos. Sci. Lett., 2, 114, doi: 10.1006/asle.2001.0044. Luterbacher, J., Xoplaki, E., Dietrich, D., Rickli, R., Jacobeit, J., Beck, C., Gyalistras, D., Schmutz, C., & Wanner, H. (2002b). Reconstruction of sea level pressure fields over the
Mediterranean Climate Variability over the Last Centuries: A Review
139
Eastern North Atlantic and Europe back to 1500. Clim. Dynam., 18, 545, doi: 10.1007/ s00382-001-0196-6. Luterbacher, J., Dietrich, D., Xoplaki, E., Grosjean, M., & Wanner, H. (2004). European seasonal and annual temperature variability, trends, and extremes since 1500. Science, 303, 1499, doi: 10.1126/science.1093877. Mangini, A., Spo¨tl, C., & Verdes, P., (2005). Reconstruction of temperature in the Central Alps during the past 2000 yr from a 18O stalagmite record. Earth Planet Sc. Let., 235, 741, doi: 10.1016/j.epsl.2005.05.010. Mann (Abbe´) (1790), Ueber die allma¨hlichen Vera¨nderungen der Temperatur und des Bodens in verschiedenen Climaten, nebst Untersuchungen u¨ber die Ursachen dieser Vera¨nderungen, Historia et Commentationes Academiae Theodoro-Palatinat, Vol. 6, Physicum Mannheimii, pp. 82–111 (reprinted in abbreviated form 1790 in Journal der Physik (Gren’s Journal ) 2, 231. Mann, M. E. (2002a). The value of multiple proxies. Science, 297, 1481. Mann, M. E. (2002b). Large-scale climate variability and connections with the Middle East in past centuries. Climatic Change, 55, 287. Mann, M. E. (2004). On smoothing potentially non-stationary climate time series. Geophys. Res. Lett., 31, 07214, doi: 10.1029/2004GL019569. Mann, M. E., Bradley, R. S., & Hughes, M. K. (1998). Global-scale temperature patterns and climate forcing over the past six centuries. Nature, 392, 779. Mann, M. E., Bradley, R. S., & Hughes, M. K. (1999). Northern Hemisphere temperatures during the past millennium: inferences, uncertainties, and limitations. Geophys. Res. Lett., 26, 759. Mann, M. E., Gille, E., Bradley, R. S., Hughes, M. K., Overpeck, J., Keimig, F. T., & Gross, W. (2000). Global temperature patterns in past centuries: an interactive presentation. Earth Interactions, 4, 1. Mann, M. E., Rutherford, S., Bradley, R. S., Hughes, M. K., & Keimig, F. T. (2003). Optimal surface temperature reconstructions using terrestrial borehole data. J. Geophys. Res., 108, doi:10.1029/202JD002532. Mann, M. E., Rutherford, S., Wahl, E. R., & Ammann, C. M. (2005). Testing the fidelity of methodologies used in proxy-based reconstructions of past climate, J. Climate, 18, 4097. Mann, M.E., & Rutherford, S. (2002). Climate reconstruction using ‘Pseudoproxies’, Geophys. Res. Lett., 29(10), 1501, doi: 10.1029/2001GL014554. Martin-Vide, J., & Barriendos, M. (1995). The use of rogation ceremony records in climatic reconstruction: a case study from Catalonia (Spain). Climatic Change, 30, 201. Maugeri, M., Buffoni, L., & Chlistovsky, F. (2002a). Daily Milan temperature and pressure series (1763-1998): history of the observations and data and metadata recovery. Climatic Change, 53, 101. Maugeri, M, Buffoni, L., Delmonte, B., & Fassina, A. (2002b). Daily Milan temperature and pressure series (1763-1998): completing and homogenising the data. Climatic Change, 53, 119. Maugeri, M., Brunetti, M., Monti, F., & Nanni, T. (2004). Sea-level pressure variability in the Po Plain (1765–2000) from homogenized daily secular records. Int. J. Climatol., 24, 437. Maury, M. F. (1854). Maritime Conference held in Brussels for devising a uniform system of meteorological observations at sea, Explanations and Sailing directions to Accompany the Wind and Current Charts 6th ed. Washington D. C.
140 Mediterranean Climate Variability
McDermott, F., Mattey, D. P., & Hawkesworth, C. J. (2001). Centennial scale Holocene climate variability revealed by a high-resolution speleothem 18O record from S. W. Ireland. Science, 294, 1328. Meko, D. M. (1985). Temporal and spatial variation of drought in Morocco: drought, water management and food production. Conference Proceedings, Agadir (Morocco), 21–24 November 1985. Menzel, A. (2005). A 500 year pheno-climatological view on the 2003 heatwave in Europe assessed by grape harvest dates. Meteorol. Z., 14, 75. Milly, P. C. D., Wetherald, R. T., Dunne, K. A., & Delworth, T. L. (2002). Increasing risk of great floods in a changing climate. Nature, 415, 514. Min, S.-K., Legutke, S., Hense, A., & Kwon, W.-T. (2002). Internal variability in a 1000-yr control simulation with the coupled climate model ECHO-G – I. Near-surface temperature, precipitation and mean sea level pressure. Tellus, A57, 605. Mitchell, T. D., & Jones, P. D. (2005). An improved method of constructing a database of monthly climate observations and associated high-resolution grids. Int. J. Climatol., 25, 693. Mitchell, T. D., Carter, T. R., Jones, P. D., Hulme, M., & New, M. (2004). A comprehensive set of high-resolution grids of monthly climate for Europe and the globe: the observed record (1901–2000) and 16 scenarios (2001–2100). Tyndall Centre Working Paper 55. Moberg, A. D., Sonechkin, M., Holmgren, K., Datsenko, N. M., & Karlen, W. (2005). Highly variable Northern Hemisphere temperatures reconstructed from low- and high-resolution proxy data. Nature, 433, 613, doi: 10.1038/nature03265. Montagna, P., McCulloch, M., Mazzoli, C., & Silenzi, S. (2004). B/Ca, Sr/Ca, U/Ca and Mg/Ca ratios of a non-tropical coral (Cladocora caespitosa) from the Northern Adriatic Sea (Mediterranean Sea) and their relationship to sea surface temperature. 10th International Coral Reef Symposium, June 28-July 2 2004, Okinawa, Japan. Montagna, P., McCulloch, M., Taviani, M., Remia, A., & Rouse, G. (2005a). High resolution trace and minor element compositions in deep-sea Azooxanthellate solitary corals (Desmophyllum dianthus) from the Mediterranean Sea and the Great Australia Bight: cold-water corals and ecosystems. In: A. Freiwald, & J. M. Roberts (Eds), Cold-water Corals and Ecosystems (Springer, Berlin Heidelberg, 1109). Montagna, P., McCulloch, M., Taviani, M., Remia, A., & Mazzoli, C. (2005b). In-situ high resolution minor and trace element compositions in Desmophyllum dianthus from the Mediterranean Sea and the Pacific Ocean. 3rd International Symposium on Deep-Sea Corals Science and Management, November 28-December 2, 2005, Miami, USA. Monterosato, M. T. A., (1892), Monografia dei vermeti del Mediterraneo. Bullettino della Societa Malacologica Italiana, 17, 7. Mottaghy, G., & Rath, V. (2005). Latent heat effects in subsurface heat transport modeling and their impact on paleotemperature reconstructions. Geophys. J. Int., in press. Munaut, A. V. (1982). The Mediterranean area: climate from tree-rings. In: M. K. Hughes, P. M. Kelly, J. M. Pilcher, and V. C. LaMarche (Eds), (Cambridge University Press, Cambridge). Nitoiu, D., & Beltrami, H. (2005). Subsurface thermal effects of land use changes. J. Geophys. Res., 110, F01005, doi: 10.1029/2004JF000151.
Mediterranean Climate Variability over the Last Centuries: A Review
141
Overpeck, J., & 17 coauthors (1997). Arctic environmental change of the last four centuries. Science, 278, 1251. Paeth, H., & Hense, A. (2005). Mean versus extreme climate in the Mediterranean region and its sensitivity to future global warming conditions, Meteorol. Z., 14, 329. Pal, J. S., Giorgi, F., & Bi, X. (2004). Consistency of recent European summer precipitation trends and extremes with future regional climate projections. Geophys. Res. Lett., 31, L13202, doi: 10.1029/2004GL019836. Palmer, W. C. (1965). Meteorological drought. Research Paper No. 45, U.S. Department of Commerce Weather Bureau, Washington, D.C. Palmer, T. N., & Ra¨isa¨nen, J. (2002). Quantifying the risk of extreme seasonal precipitation events in a changing climate. Nature, 415, 512. Park, J.-S., Jung, H.-S., Kim, R.-S., & Oh, J.-H. (2001). Modelling summer extreme rainfall over the Korean peninsula using Wakeby distribution. Int. J. Climatol., 21, 1371. Pasquale, V., Verdoya, M., Chiozzi, P., & Safanda, J. (2000). Evidence of climate warming from underground temperatures in NW Italy. Global Planet. Change, 25, 215. Pasquale, V., Verdoya, M., Chiozzi, P., Bodri, L., & Bellani, S. (2005). Temperature signal in the underground for climate history reconstruction in Italy. Global Planet. Change, 47, 36, doi:10.1016/j.gloplacha.2004.11.015. Pauling, A., Luterbacher, J., & Wanner, H. (2003). Evaluation of proxies for European and North Atlantic temperature field reconstructions. Geophys. Res. Lett., 30, 1787, doi: 10.1029/2003GL017589. Pauling, A., Luterbacher, J., Casty, C., & Wanner, H. (2005). 500 years of gridded highresolution precipitation reconstructions over Europe and the connection to large-scale circulation, Clim. Dynam., doi: 10.1007/S00382-005-0090-8. Paulsen, D. E., Li, H.-C., & Ku, T.-L. (2003). Climate variability in central China over the last 1270 years revealed by high-resolution stalagmite records. Quaternary Sci. Rev., 22, 691. Pfister, C. (1992). Monthly temperature and precipitation patterns in Central Europe from 1525 to the present. In: R. S. Bradley, and P. D. Jones (Eds), A methodology for quantifying man-made evidence on weather and climate, (Routledge, London and New York, pp. 118–142). Pfister, C. (1999). Wetternachhersage. 500 Jahre Klimavariationen und Naturkatastrophen. Haupt Verlag, Bern. Pfister, C., & Bra´zdil, R. (1999). Climatic variability in sixteenth-century Europe and its social dimension: a synthesis. Climatic Change, 43, 5. Pfister, C., Luterbacher, J., Schwarz-Zanetti, G., & Wegmann, M. (1998). Winter air temperature variations in Western Europe during the Early and High Middle Ages (AD 750–1300). Holocene, 8, 535. Pfister, C., Bra´zdil, R., Glaser, R., Barriendos, M., Camuffo, D., Deutsch, M., Dobrovolny, P., Enzi, S., Guidoboni, E., Kotyza, O., Militzer, S., Ra´cz, L., & Rodrigo, F. S. (1999). Documentary evidence on climate in sixteenth-century Europe. Climatic Change, 43, 55. Pfister, C., Bra´zdil, R., & Barriendos, M. (2002). Reconstructing past climate and natural disasters in Europe using documentary evidence: PAGES (Past Global Changes) newsletter on documentary evidence, C. Pfister, H. Wanner, C. Kull, & K. Alverson (Eds) Vol. 10, No 3, 6. Available through: http://www.pages-igbp.org/. Pfister, C. (2005). Weeping in the snow, the second period of Little Ice Age-type Impacts. In: W. Behringer, Lehmann, H. & Pfister, C. (Eds), Cultural consequences of the ‘Little
142 Mediterranean Climate Variability
Ice Age’ (‘Kulturelle Konsequenzen der Kleinen Eiszeit’) (Vandenhoech & Ruprecht, Go¨ttingen, pp. 31–86). Pichard, G. (1999). Espace et nature en Provence du XVI au XVIII sie`cle. Universite´ d’Aix-Marseille I, Aix-en-Provence. Piervitali, E., & Colacino, M. (2001). Evidence of drought in western Sicily during the period 1565–1915 from liturgical offices. Climatic Change, 49, 225. Pirazzoli, P. A., & Montaggioni, L. F. (1989). Crustal block movements from Holocene shorelines: Rhodes Island Greece. Tectonophysics, 170, 89. Pirazzoli, P. A., Laborel, J., & Stiros, S. C. (1996). Earthquake clustering in the Eastern Mediterranean during historical times. J. Geophys. Res., 101, 6083. Pla, S., & Catalan, J. (2005). Chrysophyte cysts from lake sediments reveal the submillennial winter/spring climate variability in the northwestern Mediterranean region throughout the Holocene. Clim. Dynam., 14, 263, doi: 10.1007/s00382-004-0482-1. Pollack, H. N., & Smerdon, J. E. (2004). Borehole climate reconstructions: spatial structure and hemispheric averages. J. Geophys. Res., 109, D11106, doi: 10.1029/ 2003JD004163. Pons-Branchu, E., Hillaire-Marcel, C., Deschamps, P., Ghaleb, B., & Sinclair, D. J. (2005). Early diagenesis impact on precise U-series dating of deep-sea corals: example of a 100–200-year old Lophelia pertusa sample from the northeast Atlantic. Geochim. Cosmochim. Ac., 69, 4865. Prohom, M., & Barriendos, M. (2004). Potencialidad de los diarios de navegacio´n catalanes comoinstrumento de reconstruccio´n clima´tica sobre los oce´anos (ss XVIII a XX). Boletı´n RECLIDO 1(13). Available at www.ucm.es/info/reclido. Przybylak, R., Majorowicz, J., Wo´jcik, G., Zielski, A., Chorazyczewski, W., Marciniak, K., Nowosad, W., Olin´ski, P., & Syta, K. (2005). Temperature changes in Poland from the 16th to the 20th centuries. Int. J. Climatol., 25, 773. Purgstall, J., & von H. (1983). Ottoman state history, vol 1–7 (in Turkish). Translator, Vecdi Bu¨ru¨n, U¨cdal, Istanbul. Qian, W.-H., & Zhu, Y.-F. (2002). Little Ice Age climate near Beijing, China, inferred from historical and stalagmite records. Quaternary Res., 57, 109. Qian, W., Hu, Q., Zhu, Y., & Lee, D.-K. (2003a). Centennial-scale dry-wet variations in East Asia. Clim. Dynam., 21, 77. Qian, W.-H., Chen, D., Zhu, Y., & Shen, H.-Y. (2003b). Temporal and spatial variability of dryness/wetness in China during the last 530 years. Theor. Appl. Climatol., 76, 13, doi: 10.1007/s00704-003-0009-4. Ra´cz, L. (1999). Climate History of Hungary since 16th century: past, present and future. Pa´l, Pe´cs, p 160, revised and updated version of September 2001. Raible, C. C., Casty, C., Luterbacher, J., Pauling, A., Esper, J., Frank, D. C., Bu¨ntgen, U., Roesch, A. C., Tschuck, P., Wild, M., Vidale, P-L., Scha¨r, C., & Wanner, H. (2005). Climate variability - observations, reconstructions, and model simulations for the Atlantic-European and Alpine region from 1500–2100 AD, Climatic Change, in press. Rajver, D., Safanda, J., & Shen, P. Y. (1998). The climate record innverted from borehole temperatures in Slovenia. Tectonophysics, 291, 263. Rebetez, M. (2004). Summer 2003 maximum and minimum daily temperatures over a 3300 m altitudinal range in the Alps. Clim. Res., 27, 45. Reiter, M. (2005). Possible ambiguities in subsurface temperature logs – consideration of groundwater flow and ground surface temperature change. Pure Appl. Geophys., 162, 343.
Mediterranean Climate Variability over the Last Centuries: A Review
143
Repapis, C. C., Schuurmans, C. J. E., Zerefos, C., & Ziomas, J. (1989). A note on the frequency of occurrence of severe winters as evidenced in monastery and historical records from Greece during the period 1200–1900 A.D. Theor. Appl. Climatol., 39, 213. Rico Sinobas, M., (1851). Memoria sobre las causas meteorologico fı´ sicas que producen las constantes sequı´ as de Murcia y Almerı´ a, sen˜alando los medios de atenuar sus efectos. D. S. Compagni, Madrid. Rimbu, N., Lohmann, G., Felis, T., & Pa¨tzold, J. (2001). Arctic Oscillation signature in a Red Sea coral. Geophys. Res. Lett., 28, 2959. Rimbu, N., Lohmann, G., Felis, T., & Pa¨tzold, J. (2003a). Shift in ENSO teleconnections recorded by a northern Red Sea coral. J. Climate, 16, 1414. Rimbu, N., Lohmann, G., Kim, J.-H., Arz, H. W., & Schneider, R. (2003b). Arctic/ North Atlantic Oscillation signature in Holocene sea surface temperature trends as obtained from alkenone data. Geophys. Res. Lett., 30, 1. Rimbu, N., Felis, T., Lohmann, G., & Patzold, J. (2005). Winter and summer climate patterns in the European-Middle East during recent centuries as documented in a northern Red Sea coral record. Holocene, in press. Rimi, A. (2000). Evidence of a recent warming in the north of Morocco from disturbed geothermal gradients. Geodinamica Acta, 1, 19. Roberts, M. S., Smart, P. L., & Baker, A. (1998). Annual trace element variations in a Holocene speleothem. Earth Planet. Sc. Lett., 154, 237. Rodrigo, F. S. (2002). Changes in climate variability and seasonal rainfall extremes: a case study from San Fernando (Spain), 1821–2000. Theor. Appl. Climatol., 72, 193. Rodrigo, F. S., Esteban-Parra, M. J., & Castro-Diez, Y. (1998). On the use of the Jesuit order private correspondence records in climate reconstructions: a case study from Castille (Spain) for 1634-1648 AD. Climatic Change, 40, 625. Rodrigo, F. S., Esteban-Parra, M. J., Pozo-Va´zquez, D., & Castro-Dı´ ez, Y. (1999). A 500-year precipitation record in Southern Spain. Int. J. Climatol., 19, 1233. Rodrigo, F. S., Esteban-Parra, M. J., Pozo-Va´zquez, D., & Castro-Dı´ ez, Y. (2000). Rainfall variability in Southern Spain on decadal to centennial time scales. Int. J. Climatol., 20, 721. Rodrigo, F. S., Esteban-Parra, M. J., Pozo-Va´zquez, D., & Castro-Dı´ ez, Y. (2001). A reconstruction of the winter North Atlantic Oscillation Index back to AD 1501 using documentary data in Southern Spain. J. Geophys. Res., 106, 14805. Rodrı´ guez, R., Barriendos, M., Jones, P. D., Martin-Vide, J., & Pena, J. C. (2001). Long pressure series for Barcelona (Spain). Daily reconstruction and monthly homogenization. Int. J. Climatol., 21, 1693. Roeckner, E., Arpe, K., Bengtsson, L., Christoph, M., Claussen, M., Du¨menil, L., Esch, M., Giorgetta, M., Schlese, U., & Schulzweida, U. (1996). The atmospheric gereneral circulation model ECHAM4: model description and simulation of present-day climate. Max-Planck-Institut fu¨r Meteorologie, Hamburg, Report No. 218. Roeckner, E., Bengtsson, L., Feichter, J., Lelieveld, J., & Rodhe, H. (1999). Transient climate change simulations with a coupled atmosphere-ocean GCM including the tropospheric sulfur cycle. J. Climate, 12, 3004. Rutherford, S., Mann, M. E., Delworth, T. L., & Stouffer, R. (2003). Climate field reconstruction under stationary and nonstationary forcing. J. Climate, 16, 462. Rutherford, S., Mann, M. E., Osborn, T. J., Bradley, R. S., Briffa, K. R., Hughes, M. K., & Jones, P. D. (2005). Proxy-based Northern Hemisphere surface temperature
144 Mediterranean Climate Variability
reconstructions: sensitivity to method, predictor network, target season, and target domain. J. Climate, 18, 2308. Ryan, P. T. (1997). Modern regression methods. John Wiley and Sons, New York. Safriel, U. (1966). Recent vermetid formation on the Mediterranean shore of Israel. Proc. of the Malacological Society of London, 37, 27. Safriel, U. (1974). Vermetid gastropods and intertidal reefs in Israel and Bermuda. Science, 186, 1113. Saz, M. A. (2004). Temperaturas y precipitaciones en la mitad norte de Espan˜a desde el siglo XV. Estudio dendroclima´tico. Consejo de Proteccio´n de la Naturaleza. Diputacio´n General de Arago´n, Zaragoza. p 294. Saz, M., & Creus-Novau, J. (1999). La variabilidad del clima espan˜ol en el pasado: frecuencia de valores extremos de temperatura y precipitacio´n reconstruidas desde el s. XV: La Climatologı´ a espan˜ola en los albores del siglo XXI, Raso and Martı´ n-Vide (Eds), Vilassar, Oikos-tau, Asociacio´n Espan˜ola de Climatologı´ a, Serie A, No 1. Saz, M. A., Creus, J., & Cuadrat, J. M. (2003). Mean summer temperatures dendroclimatic reconstructions in northeast of Spain. Comparison with other regional studies. Geophys. Res., 5. Scha¨r, C., Vidale, P. L., Lu¨thi, D., Frei, C., Ha¨berli, C., Liniger, M. A., & Appenzeller, C. (2004). The role of increasing temperature variability in European summer heatwaves. Nature, 427, 332. Schilman, B., Bar-Matthews, M., Almogi-Labin, A., & Luz, B. (2001). Global climate instability reflected by Eastern Mediterranean marine records during the late Holocene. Palaeogeogr. Palaeoecol., 176, 157. Schmidt, G. A., Shindell, D. T., Miller, R. L., Mann, M. E., & Rind, D. (2004). General circulation modelling of Holocene climate variability. Quarternary Sci. Rev., 23, 2167. Scho¨nwiese, C. D., Staeger, T., & Tromel, S. (2004). The hot summer 2003 in Germany. Some preliminary results of a statistical time series analysis. Meteorol. Z., 13, 323. Schwarcz, H. P. (1986). Geochronology and isotope geochemistry in speleothems. Handbook of environmental isotope geochemistry Elsevier, Amsterdam. Serre-Bachet, F., & Guiot, J., (1987). Summer temperature changes from tree-rings in the Mediterranean area during the last 800 years: abrupt climatic change, evidences and implications. Berger, W. & Labeyrie, L. (Eds) (Reidel, Dordrecht). Serre-Bachet, F., Guiot, J., & Tessier, L. (1992). Dendroclimatic evidence from southwestern Europe and northwestern Africa: climate since AD 1500. Routledge, London. Shabalova, M. V., & van Engelen, A. F. V. (2003). Evaluation of a reconstruction of winter and summer temperatures in the Low Countries, AD 764–1998. Climatic Change, 58, 219. Shen, P. Y., & Beck, A. E. (1991). Least squares inversion of borehole temperature measurements in functional space. J. Geophys. Res., 96, 19965. Sheppard, P. R., Tarasov, P. E., Graumlich, L. J., Heussner, K.-U., Wagner, M., O¨sterle, H., & Thompson, L. G. (2005). Annual precipitation since 515 BC reconstructed from living and fossil juniper growth of northeastern Qinghai Province, China. Clim. Dynam., 23, 869. Shindell, D. T., Schmidt, G. A., Mann, M. E., Rind, D., & Waple, A. (2001a). Solar forcing of regional climate change during the Maunder Minimum. Science, 294, 2149. Shindell, D. T., Schmidt, G. A., Miller, R. L., & Rind, D. (2001b). Northern Hemisphere winter climate response to greenhouse gas, ozone, solar, and volcanic forcing. J. Geophys. Res., 106, 7193.
Mediterranean Climate Variability over the Last Centuries: A Review
145
Shindell, D. T., Schmidt, G. A., Miller, R. L., & Mann, M. E. (2003). Volcanic and solar forcing of climate change during the preindustrial era. J. Climate, 16, 4094. Shindell, D. T., Schmidt, G. A., Mann, M. E., & Faluvegi, G. (2004). Dynamic winter climate response to large tropical volcanic eruptions since 1600. J. Geophys. Res. 109. Silenzi, S., Antonioli, F., & Chemello, R. (2004). A new marker for sea surface temperature trend during the last centuries in temperate areas: Vermetid reef. Global Planet. Change, 40, 105. Silenzi, S., Bard, E., Montagna, P., & Antonioli, F. (2005). Isotopic and elemental records in a non-tropical coral (Cladocora caespitosa): discovery of a new highresolution climate archive for the Mediterranean Sea. Global Planet. Change, 49, 94. Song, J. (1998). Changes in dryness/wetness in China during the last 529 years. Int. J. Climatol., 18, 1345. Song, J. (2000). Reconstruction of the southern oscillation from dryness/wetness in China for the last 500 years. Int. J. Climatol., 20, 1003. Souriau, A., & Yiou, P. (2001). Grape harvest dates for checking NAO paleoreconstructions. Geophys. Res. Lett., 28, 3895. Spo¨tl, C., & Mangini, A. (2002). Stalagmite from the Austrian Alps reveal DansgardOeschger events during isotope stage 3: implications for the absolute chronology of Greenland ice cores. Earth Planet. Sc. Lett., 203, 507. Stendel, M., Morgensen, I. A., & Christensen, J. H. (2005). Influence of various forcings on global climate in historical times using a coupled AOGCM. Clim. Dyn., 26, 1. Stephenson, T. A., & Stephenson, A. (1954). The Bermuda Islands. Endeavour, 50, 72. Stott, P. A., Stone, D. A., & Allen, M. R. (2004). Human contribution to the European heatwave of 2003. Nature, 432, 610, doi: 10.1038/nature03089. Taborda, J. P., Alcoforado, M. J., & Garcia, J. C. (2004). The climate of Southern Portugal during the 18th century: reconstruction based on descriptive and instrumental sources, Geo-Ecologia, Rel.2, Centro de Estudos Geogra´ficos, Lisboa (in Portuguese with English summary). Taviani, M., Freiwald, A., & Zibrowius, H. (2005). Deep coral growth in the Mediterranean Sea: an overview. In: A. Freiwald, and J. M. Roberts (Eds), Cold-water Corals and Ecosystems (Springer, Berlin, Heidelberg). Telelis, I., (2000). Medieval Warm Period and the beginning of the Little Ice Age in Eastern Mediterranean. An approach of physical and anthropogenic evidence: Byzanz als Raum. Zu Methoden und Inhalten der historischen Geographie des O¨stlichen Mittelmeerraumes, Vero¨ffentlichung der Kommission fu¨r die Tabula Imperii Byzantini, In: K. Belke, F. Hild, J. Koder, & P. Soustal (Eds), Denkschrift 7. Telelis, I., (2004). Meteorological phenomena and climate in Byzantium: approach of sources’ information and empirical indications concerning climatic fluctuations in Eastern Mediterranean and the Middle East (AD 300–1500). Academy of Athens, Ponimata No 5, Athens (in Greek with English summary). Tett, S. F. B., Betts, R., Crowley, T. J., Gregory, J., Johns, T. C., Jones, A., Osborn, T. J., O¨stro¨m, E., Roberts, D. L., & Woodage, M. J. (2005). The impact of natural and anthropogenic forcings on climate and hydrology since 1550. Clim. Dynam., submitted. Till, C., & Guiot, J. (1990). Reconstruction of precipitation in Morocco since 1100 AD based on Cedrus atlantica tree-ring widths. Quaternary Res., 33, 337. Torrence, C., & Compo, G. P. (1998). A practical guide to wavelet analysis. B. Am. Meteorol. Soc., 79, 61.
146 Mediterranean Climate Variability
Touchan, R., & Hughes, M. K. (1999). Dendrochronology in Jordan. J. Arid Environ., 42, 291. Touchan, R., Meko, D., & Hughes, M. K. (1999). A 396-year reconstruction of precipitation in Southern Jordan. J. Americ. Wat. Res. Ass., 35, 49. Touchan, R., Garfin, G. M., Meko, D. M., Funkhouser, G., Erkan, N., Hughes, M. K., & Wallin, B. S. (2003). Preliminary reconstructions of spring precipitation in southwestern Turkey from tree-ring width. Int. J. Climatol., 23, 157. Touchan, R., Funkhouser, G., Hughes, M. K., & Erkan, N. (2005a). Standardized precipitation indices reconstructed from tree-ring width for the Turkish region, Climatic Change, 72, 339. Touchan, R., Xoplaki, E., Funkhouser, G., Luterbacher, J., Hughes, M. K., Erkan, N., Akkemik, U¨., & Stephan, J. (2005b). Reconstructions of spring/summer precipitation for the Eastern Mediterranean from tree-ring widths and its connection to large-scale atmospheric circulation, Clim. Dyn., 25, 75, doi: 10.1007/ s00382-005-0016-5. Treble, P., Shelley, J. M. G., & Chappell, J. (2003). Comparison of high resolution sub-annual records of trace elements in a modern (1911–1992) speleothem with instrumental climate data from southwest Australia. Earth Planet. Sc. Lett., 216, 141. Treble, P., Chappell, J., & Shelley, J. M. G. (2005). Complex speleothem growth processes revealed by trace element mapping and scanning electron microscopy of annual layers. Geochim. Cosmochim. Ac., 69, 4855. Trenberth, K., & Paolino, D. A. (1980). The Northern Hemisphere sea-level pressure data set: trends, errors and discontinuities. Mon. Wea. Rev., 108, 855. Trigo, R. M., Garcı´ a-Herrera, R., Dı´ az, J., Franco Trigo, I., & Anto´nia Valente, M. (2005). How exceptional was the early August 2003 heatwave in France? Geophys. Res. Lett., 32, doi:10.1029/2005GL022410. van Andel, T., & Laborel, J. (1964). Recent high relative sea level stand near Recife, Brazil. Science, 145, 580. van der Schrier, G., & Barkmeijer, J. (2005). Bjerknes hypothesis on the coldness during AD 1790–1820 revisited. Clim. Dynam., 24, 355. van Engelen, A. F. V., Buisman, J., & Ijnsen, F. (2001). A millennium of weather, winds and water in the low countries. In: P. D. Jones et al. (Eds), History and Climate: Memories of the Future, (Kluwer Academic Press, New York, Boston, London, pp. 101–124). Vinther, B. M., Johnsen, S. J., Andersen, K. K., Clausen, H. B., & Hansen, A. W. (2003a). NAO signal recorded in the stable isotopes of Greenland ice cores. Geophys. Res. Lett., 30, 1387, doi:10.1029/2002GL016193. Vinther, B. M., Andersen, K. K., Hansen, A. W., Schmith, T., & Jones, P. D. (2003b). Improving the Gibraltar/Reykjavik NAO Index. Geophys. Res. Lett., 30, 2222, doi:10.1029/2003GL018220. von Storch, H., Zorita, E., Jones, J., Dimitriev, Y., Gonza´lez-Rouco, F., & Tett, S. (2004). Reconstructing past climate from noisy data. Science, 306, 679. Wagner, S., & Zorita, E., (2005). The influence of volcanic, solar and CO2 forcing on the temperatures in the Dalton Minimum (1790–1830): a model study, Clim. Dyn., 25, 205, doi:10.1007/s00382-005-0029-0. Wang, W. C. (2002). Workshop on historical climate reconstruction over East Asia, Session: 2, Beijing, China, October 14–16, 2002.
Mediterranean Climate Variability over the Last Centuries: A Review
147
Wang, S.-W., Gong, D., & Zhu, J. (2001). Twentieth-century climatic warming in China in the context of the Holocene. Holocene, 11, 313. Wang, S.-W., & Zhao, Z.-C. (1981). Droughts and floods in China, 1470–1979. In: T. M. L. Wigley, M. J. Ingram, and G. Farmer (Eds), Climate and History (Cambridge Univ. Press, New York). Waple, A. M., Mann, M. E., & Bradley, R. S. (2002). Long-term patterns of solar irradiance forcing in model experiments and proxy based surface temperature reconstructions. Clim. Dynam., 18, 563. Wells, N., Goddard, S., & Hayes, M. J. (2004). A self-calibrating Palmer Drought Severity Index. J. Climate, 17, 2335. Wheeler, D. A. (1987). The Trafalgar storm: 22–29 October 1805. Meteorol. Mag., 116, 197. Wheeler, D. A. (2001). The weather of the European Atlantic Seaboard during October 1805: an exercise in historical climatology. Climatic Change, 48, 361. Wheeler D. A., (2005). An examination of the accuracy and consistency ships’ logbook weather observations and records, Climatic Change, 73, 97. Wheeler, D. A., & Sua´rez Domı´ nguez, J. M. (2006). Climatic reconstructions for the north-east Atlantic region AD 1685–1700: a new source of evidence from naval logbooks, Holocene, 16, 39. Wolff, J., Maier-Raimer, O., & Legutke, S. (1997). The Hamburg ocean primitive equation model, Thechnical report. No. 13, German Climate Computer Center (DKRZ), Hamburg. Worley, S. J., Woodruff, S. D., Reynolds, R. W., Lott, N., & Lubker, S. J. (2005). ICOADS Release 2.1 data and products. Int. J. Climatol., 25, 823. Xoplaki, E., (2002). Climate variability over the Mediterranean. Ph.D. thesis, University of Bern, Switzerland. Available at http://sinus.unibe.ch/klimet/docs/ phd_xoplaki.pdf. Xoplaki, E., Maheras, P., & 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. Xoplaki, E., Gonza´lez-Rouco, J. F., Luterbacher, J., & Wanner, H. (2003). Mediterranean summer air temperature variability and its connection to the large-scale atmospheric circulation and SSTs. Clim. Dynam., 20, 723, doi: 10.1007/ s00382-003-0304-x. Xoplaki, E., Gonza´lez-Rouco, J. F., Luterbacher, J., & Wanner, H. (2004). Wet season Mediterranean precipitation variability: influence of large-scale dynamics and trends. Clim. Dynam., 23, 63, doi: 10.1007/s00382-004-0422-0. Xoplaki, E., Luterbacher, J., Paeth, H., Dietrich, D., Steiner, N., Grosjean, M., & Wanner, H., (2005). European spring & autumn temperature variability and change of extremes over the last half millennium, Geophys. Res. Lett., 32, L15713, doi: 10.1029/ 2005GL023424. Yang, B., Braeuning, A., Johnson, K. R., & Yafeng, S. (2002). General characteristics of temperature variation in China during the last two millennia. Geophys. Res. Lett., 29, doi: 10.1029/2001GL014485. Yoshimori, M., Stocker, T. F., Raible, C. C., & Renold, M. (2005). Externally-forced and Internal Variability in ensemble climate simulations of the Maunder Minimum, J. Climate, 18, 4253. Zerefos., C. S., Tourpali, K., & Repapis, C. (2002). Global anomaly patterns associated with extreme summer and winter air temperature and precipitation in Athens, Greece.
148 Mediterranean Climate Variability
Workshop on historical climate reconstruction over East Asia, Session: 2, Beijing, China, October 14–16, 2002. Zhang, J., & Crowley, T. J. (1989). Historical climate records in China and reconstruction of past climates. J. Climate, 2, 833. Zhang, X., Zwiers, F. W., & Li, G. (2004). Monte Carlo experiments on the detection of trends in extreme values. J. Climate, 17, 1945. Zhang, Z., Mann, M. E., & Cook, E. R. (2005). Alternative methods of proxy-based climate field reconstruction: application to summer drought over the conterminous United States back to AD 1700 from tree-ring data. Holocene, 14, 502. Zorita, E., & Gonza´lez-Rouco, J. F. (2002). Are temperature sensitive proxies adequate for North Atlantic Oscillation reconstructions? Geophys. Res. Lett., 29, 14, doi: 10.1029/2002GL015404. Zorita, E., Gonza´lez-Rouco, J. F., & Legutke, S. (2003). Statistical temperature reconstruction in a 1000-year-long control climate simulation an exercise with Mann’s et al. (1998) method. J. Climate, 16, 1378. Zorita, E., von Storch, H., Gonza´lez-Rouco, J. F., Cubasch, U., Luterbacher, J., Fischer-Bruns, I., Legutke, S., & Schlese, U. (2004). Climate evolution in the last five centuries simulated by an atmosphere-ocean model: global temperatures, the North Atlantic Oscillation and the Late Maunder Minimum. Meteorol. Z., 13, 271. Zorita, E., Gonza´lez-Rouco, J. F., von Storch, H., Monta´vez, J. P., & Valero, F. (2005). Natural and anthropogenic modes of surface temperature variations in the last thousand years. Geophys. Res. Lett., 32, L08707, doi:10.1029/2004GL021563.
Chapter 2
Relations between Climate Variability in the Mediterranean Region and the Tropics: ENSO, South Asian and African Monsoons, Hurricanes and Saharan Dust Pinhas Alpert,1 Marina Baldi,2 Ronny Ilani,1 Shimon Krichak,1 Colin Price,1 Xavier Rodo´,3 Hadas Saaroni,1 Baruch Ziv,4 Pavel Kishcha,1 Joseph Barkan,1 Annarita Mariotti5 and Eleni Xoplaki6 1
Tel Aviv University, Israel (
[email protected],
[email protected],
[email protected],
[email protected],
[email protected],
[email protected]) 2 IBIMET – CNR, Italy (
[email protected]) 3 ICREA and Climate Research Laboratory, PCB-Univ. of Barcelona, Spain (
[email protected]) 4 The Open University of Israel (
[email protected]) 5 Earth System Science Interdisciplinary Center, USA and ENEA, Italy (
[email protected]) 6 Institute of Geography and NCCR Climate, University of Bern, Switzerland (
[email protected])
2.1. Introduction The Mediterranean climate is affected by several tropical and subtropical systems as illustrated by some evidence presented in this chapter. These factors range from the El Nin˜o Southern Oscillation (ENSO) and tropical hurricanes to the South Asian Monsoon and Saharan dust. This leads to complex features in the Mediterranean climate variability. In the following sections, we review some tropical and subtropical teleconnections to the Mediterranean climate in the following order: El Nin˜o Southern Oscillation is elaborated in Section 2.2, the South Asian Monsoon is discussed in Section 2.3, Section 2.4 is dedicated to African monsoon, tropical cyclones are discussed in Section 2.5 and finally Red Sea Trough intrusions into the Eastern Mediterranean and the Saharan dust are discussed respectively in the Sections 2.6 and 2.7.
150 Mediterranean Climate Variability
2.2. ENSO Impact on the Mediterranean Climate 1
The El Nin˜o Southern Oscillation (ENSO) phenomenon is recognized as a major source for global climate variability (Halpert and Ropelewski, 1992) either via standing modes over the entire Tropics or via coherent large-scale low-frequency spatial patterns referred to as ‘‘teleconnections’’ over midlatitudes (see Wallace and Gutzler, 1981; Allan et al., 1996; Diaz et al., 2001 for reviews on ENSO). Several studies have dealt with the underlying physics of the phenomenon and with the worldwide implications for climate (e.g. van Loon and Madden, 1981; Kiladis and Diaz, 1989; Ropelewski and Halpert, 1987; Trenberth et al., 1998; Diaz et al., 2001). The impact of ENSO on the climate of extra-tropical regions, as well as the mechanism responsible for anomalies in the tropical Pacific sea surface temperatures (SST) having worldwide impacts are poorly understood and documented (Pozo-Va´zquez et al., 2001). The El Nin˜o phenomenon is related to the warming of the eastern Pacific sea surface temperatures (SST) for an extended period of 6–12 months, and sometimes longer. The SST distribution is directly linked to the atmospheric pressure patterns over the Pacific, with a low pressure cell being located above the warm pool in the western Pacific during normal conditions, while moving eastward with the warm pool in El Nin˜o years. The atmospheric pressure oscillation between the west and central Pacific is known as the Southern Oscillation (SO). Positive pressure anomalies over Australia and Indonesia are associated with the warm El Nin˜o conditions in the eastern Pacific, while negative pressure anomalies over Australia are associated with the cold La Nin˜a conditions in the eastern Pacific. Due to the strong link between the SSTs and the atmospheric pressure, the phenomenon is often referred to as the El Nin˜o/Southern Oscillation. During warm (El Nin˜o) episodes, the normal patterns of tropical precipitation and atmospheric circulation become disrupted. The abnormally warm waters in the equatorial central and eastern Pacific give rise to enhanced cloudiness and rainfall in that region, especially during the boreal winter and spring seasons. At the same time, rainfall is reduced over Indonesia, Malaysia and northern Australia. Thus, the normal Walker Circulation during winter and spring, which features rising air, cloudiness and rainfall over the region of Indonesia and the western Pacific, and sinking air over the equatorial eastern Pacific, becomes weaker than normal, and for strong warm episodes, it may actually reverse. The increased heating of the tropical atmosphere over the central and eastern Pacific during warm episodes affects global atmospheric circulation features, such as the jet streams in the subtropics and in the temperate latitudes of the 1 Much of the preface of this section is based on Xoplaki (2002) and the Climate Prediction Center (CPC) website http://www.cpc.ncep.noaa.gov/
Climate Variability in the Mediterranean Region and the Tropics
151
winter hemisphere. The jet streams over the eastern Pacific Ocean are stronger than normal during warm episodes. Also, during warm episodes, extra-tropical storms and frontal systems follow paths that are significantly different from normal, resulting in persistent temperature and precipitation anomalies in many regions. Significant departures from normal conditions, for the Northern Hemisphere (NH) winter and summer seasons, can be found at the Climate Prediction Center (CPC) site: http://www.cpc.ncep.noaa.gov/products/analysis_ monitoring/lanina/index.html). It has been proposed that ENSO exerts a positive forcing on tropical North Atlantic SSTs and this effect is strongest in boreal spring (Enfield and Mayer, 1997). However, it has been argued that only when tropical SST anomalies are large (strong ENSO events), the ENSO signal can be found in the extra-tropics (Huang et al., 1998; Trenberth et al., 1998). On the other hand, tropical forcing is stronger during the northern winter, coinciding with the mature stage of El Nin˜o events (Trenberth et al., 1998). It appears that the possible influence of ENSO in the North Atlantic-European area is more likely to be found during extreme events of ENSO and during the winter (Pozo-Va´zquez et al., 2001). The perturbation can be propagated downstream, as a wave train, to other longitudes in the form of Rossby waves, eventually affecting locations far away from the Pacific, particularly the North Atlantic region. Several papers have related ENSO to weather and climate variability over Europe and Africa as well as over specific countries at the Mediterranean Sea (e.g. Fraedrich and Mu¨ller, 1992; Fraedrich, 1994; Rodo´ et al., 1997; Laita and Grimalt, 1997; Moron and Ward, 1998; Rodriguez-Puebla et al., 1998; Price et al., 1998; Tu¨rkes° , 1998; Kadiog˘lu et al., 1999; Rocha, 1999; van Oldenborgh et al., 2000; Compo et al., 2001; Diaz et al., 2001; Pozo-Va´zquez et al., 2001; Giorgi, 2002; Lloyd-Hughes and Saunders, 2002). A compilation of their findings together with some others is summarized below.
2.2.1. ENSO and Eastern Mediterranean (EM) Rainfall Yakir et al. (1996) and Price et al. (1998) showed significant connections between ENSO events and winter rainfall in Israel, both indicate increased rainfall occurring in El Nin˜o winters. Price et al. (1998) also demonstrated that La Nin˜a years were associated with below normal rainfall. The 2003–2004 rainy winter in Israel, coinciding with an El Nin˜o event, supports the above. The analysis in Israel was extended to the Jordan River discharge, used as a proxy for regional rainfall, since the stream flow entering the Sea of Galilee is dominated by regional rainfall. The seasonal stream flow in the Jordan River is significantly correlated (r 0.67) with the seasonal NINO4 temperatures (Fig. 45). This implies that the tropical Pacific temperature oscillations can explain
152 Mediterranean Climate Variability
Figure 45: The winter streamflow in the Jordan River and the winter NINO4 SSTs in the tropical Pacific. Adapted from Price et al. (1998). approximately 45% of the inter-annual variability in winter rainfall in northern Israel. It is hypothesized that the reason for this strong connection is related to the position of the winter jet over the Eastern Mediterranean (EM). Israel is located at 30 N, exactly the mean latitude of the winter jet. Small shifts, in the order of 1 deg, in its mean position can have a major impact on the storm tracks, and hence on the rainfall amounts. Figure 46 shows that indeed in a composite of El-Nin˜o years, the jet over the EM moves further south by about 50–100 km. During El Nin˜o/La Nin˜a years, meridional shifts of the jet in the EM have been observed. However, the intensity of the ENSO events is not directly related to the intensity of the rainfall anomalies in Israel. This is one of the reasons the correlation coefficient is only 0.67. However El Nin˜o/La Nin˜a years have been wet/dry for 75% of the ENSO events in the last 30 years. Stream flow data in the Jordan River are only available since the end of the 1960s. However, since individual rain gauge measurements in the watershed are highly correlated (r 0.9) with the catchment’s integrated stream flow, it is possible to extend the time series back to 1922. However, the ENSO signal appears in the rainfall/streamflow data only after the mid-1970s. It is puzzling as to why these correlations are observed only in the recent record. This may be a result of the
Climate Variability in the Mediterranean Region and the Tropics
153
Figure 46: Zonal means (30 E–40 E) of west wind (m/s) in the winter period (December, January and February). The dashed lines correspond to the winters from 1982/83 to 1993/94, while the solid ones, to the El-Nino winters 1982/83, 1986/87 and 1991/92.
changes in the frequency and intensity of ENSO events since the mid-1970s. Trenberth and Hoar (1997) have shown that since the mid-1970s, there has been a significant increase in the frequency of El Nin˜o events relative to La Nin˜a events, and the intensity and period of these events has also changed. It has also been suggested that there may have been a shift in the global climate system during the 1970s, which may have resulted in a stronger Pacific-mid-latitude link during the past three decades (Wuethrich, 1995). Kadiog˘lu et al. (1999) investigated the Turkish monthly total precipitation variation at 108 meteorological stations between 1931 and 1990. They found that much of the month-to-month variability is related to El Nin˜o events. El Nin˜o events, as classified by high ENSO index, seem to produce both depressions and enhancements in the southern and northwestern parts of Turkey, respectively. During El Nin˜o years, the cyclones move towards the north. This may be the
154 Mediterranean Climate Variability
reason why there is a decreasing trend in precipitation around the southwest of Turkey (Kadiog˘lu et al., 1999).
2.2.2. ENSO and the Western Mediterranean Relationship Rodo´ et al. (1997) investigated the signatures of ENSO in Spanish precipitation and stated a coherent decrease in March/April/May following El Nin˜o events, in accordance with that stated in the pioneering studies (Ropelewski and Halpert, 1987; Kiladis and Dı´ az, 1989), and later confirmed and extended by van Oldenborgh et al. (2000) and Mariotti et al. (2002). This coherence appeared to increase in the second half of the twentieth century. Mariotti et al. (2002) also found that western Mediterranean-averaged rainfall is significantly correlated with ENSO variability during autumn, with the sign being opposite to that found in spring. A composite analysis reveals an approximate 10% increase (decrease) in seasonal rainfall for El Nin˜o (La Nin˜a) events in September/October/November, preceding the mature phase of ENSO, with an early (late) arrival of the rainy season in these regions. This relationship appears to have been stationary starting from the late 1940s (Fig. 47). Mariotti et al. (2005) investigate the Mediterranean autumn ENSO-signal in the context of the impact that ENSO events have on a larger domain extending from southwest Europe/northern Africa into parts of southwest Asia, as also found by the early work of Kiladis and Dı´ az (1989) and Mason and Goddard (2001). The observational evidence suggests a link between southwest Europe rainfall anomalies and circulation anomalies in the North Atlantic/European sector, while a more direct connection to the Indo-Pacific region and Middle-Eastern jet-stream variability for the rainfall anomalies in southwest Asia. The teleconnection mechanisms for warm and cold ENSO events appear to be different, with a prevailing signature of PNA/NAO-like variability in the former case and a more relevant role for tropical Atlantic SST anomalies in the latter (Fig. 48) (PNA stands for the Pacific Ocean–North American teleconnection pattern). Regarding ENSO signatures in the North Atlantic/European sector by using common statistical techniques, Rodo´ (2001) highlighted the difficulty in isolating ENSO signals mainly due to their spiky nature with respect to the dominating mid-latitude dynamics. Their importance for the Mediterranean climate might be high, though only for selected intervals and vanish elsewhere. Rodo´ (2001) showed this occurrence for SST anomalies in the western Mediterranean basin. The possibility of an ENSO influence through perturbations of the Atlantic Walker circulation was also highlighted by Rodo´ (2001), who stated the importance of a weak Atlantic Hadley cell as a response to
Climate Variability in the Mediterranean Region and the Tropics
155
Figure 47: Correlations between western Mediterranean rainfall (from the data base of the Climate Research Unit (CRU), University of East Anglia (UK) and Nino3.4 indices in autumn (SON, black) and spring (MAM, grey). Each value refers to the correlation in a 20-year window centered at the symbol. Full symbols are for values at least 95% significant (After Mariotti et al., 2002, Fig. 6 therein).
Figure 48: Correlations between Euro-Asian autumn rainfall and Nino3.4 indices for the period 1948–2000. Shading depict region where the correlation is at least 95% significant. Data is from CRU. (After Mariotti et al., 2005, Fig. 1a therein).
156 Mediterranean Climate Variability
anomalous warming in the eastern tropical Pacific. This is in accordance with Sutton et al. (2000), Saravanan and Chang (2000) whose results suggest that a fraction of the inter-hemispheric variability in the tropical Atlantic is forced by way of a tropical atmospheric bridge (Lau and Nath, 1996; Klein et al., 1999). Correlation between ENSO and Iberian rainfall has increased in the second half of the 20th century (Rodo´ et al., 1997), but the only relevant (significant) area is confined to the eastern part of the peninsula. Later studies confirm these connections and suggested possible mechanisms responsible for those associations (Rodo´, 2001; van Oldenborgh et al., 2000; Mariotti et al., 2002), which appear to involve a typical bipolar seesaw between the Mediterranean region and northern Europe. Correlations between ENSO and Iberian rainfall are maximum in autumn before a mature El Nin˜o phase and in the spring following the El Nin˜o peak. Sign of the correlations points to an increase in autumn rainfall in the year 0 and a decrease in spring precipitation in the year þ 1. ENSO-Iberian rainfall correlations may account for up to 50% of springtime decrease in rainfall in certain areas while slightly lower values, showing a converse association with El Nin˜o, were estimated for autumn. These values mostly concentrated in the second half of the last century, a time when correlations appear to have intensified (Rodo´ et al., 1997; Mariotti et al., 2002), particularly after the 1960s. The ENSO influence appears most relevant at inter-annual timescales than the NAO effect. At inter-annual timescales the NAO effect shows no clear signature on Iberian rainfall, except for small selected areas. Conversely, ENSO accounts for half of the total annual variance in southeast Spain and parts of Morocco. The potential for future predictability needs to be further assessed in the light of the lack of current predictors for Mediterranean climate at inter-annual timescales and provided there is a sufficient time lag of some months between the two processes here involved. A gain of the inter-annual predictability potential would be mostly relevant for agricultural systems and other economic activities with the high impact on population in the Mediterranean region (Rodo´ and Comı´ n, 2000).
2.2.3. ENSO and Extreme Mediterranean Rainfall Alpert et al. (2002) calculated relative contributions of 6 daily rainfall intensity categories to the annual rainfall amounts between 1951 and 1995 over Spain, Italy, Cyprus and Israel. Both the linear and the monotone non-linear (Spearman’s) time tests show significant increases in heavy daily rainfall in spite of decreases in annual totals. For instance, torrential rainfall in Italy,
Climate Variability in the Mediterranean Region and the Tropics
157
above 128 mm/day, increased percentage wise by a factor of 4 between 1951 and 1995. It is interesting to note that the torrential rainfall peaks were observed in the El-Nin˜o years.
2.2.4. Transient and Stationary Waves Approach Previous work has shown an ENSO-impact during boreal winters, with a trough (ridge) over southern Europe during El Nin˜o (La Nin˜a) events, accompanied by more (less) cyclones reaching the Mediterranean region. That is, both the mean flow and the sub-seasonal variations of the flow are affected by ENSO. In particular, the sub-seasonal variations tend to feedback on the anomalous mean flow. However, the impact in the Atlantic and Europe, and, in particular, in the Mediterranean region, appears to be more robust during La Nin˜a events than during El Nin˜o ones. Previous work from the Interannual and Decadal Climate Variability: Scale Interaction Experiments (SINTEX) EU project (Gualdi et al., 2003) indicated that the dominating mode of interaction – resembling the NAO – is only related to La Nin˜a but not to El Nin˜o events. Further, these modes – though defined in the Atlantic and Europe – appear to be connected to the North Pacific and North America. This suggests that transient eddies are also important in ‘‘transporting’’ the ENSO-response from the latter regions to the Atlantic and Europe. This insight gained may improve the prospects of seasonal prediction in the Atlantic/European region. Modelling experiments could cope with a complex response to ENSO through the alteration of mid-latitude internal modes of variability (e.g. NAO, East Atlantic/West Russian (EATL-WRUS), etc.), in particular with respect to future scenarios (e.g. Timmermann et al., 1999).
2.2.5. Possible Coupling Mechanism of ENSO and the Mediterranean The search for the physical mechanisms that might be responsible for the connection between the tropical Pacific and the North Atlantic European region was initiated through the exploration of ENSO signatures in different regions of the tropical Atlantic. Lanzante (1996) and Enfield and Mayer (1997) explored remote forcing of the tropical Atlantic and noted a significant correlation with ENSO. They suggested that a fraction of the inter-hemispheric variability in the tropical Atlantic is forced by way of a tropical atmospheric bridge (Lau and Nath, 1996; Klein et al., 1999). Other studies have suggested such a link along a zone from 10 N to 20 N (Curtis and Hastenrath, 1995; Nobre and Shukla, 1996; Mestas-Nun˜ez and Enfield, 2001). In addition, Sutton et al. (2000) and
158 Mediterranean Climate Variability
Saravanan and Chang (2000) suggest an influence through perturbation of the Atlantic Walker circulation. This possibility was also highlighted by MestasNun˜ez and Enfield (2001) and Rodo´ (2001), who stated the importance of a weak Atlantic Hadley cell as a response to anomalous warming in the eastern tropical Pacific. Finally, Sutton et al. (2000) suggested that a variety of competing mechanisms might be responsible for the weakening of the Atlantic cell during boreal winters. Recently, Ruiz-Barradas et al. (2003) with the aid of model simulations and the NCEP–NCAR reanalysis data for the period from 1958 to 1993, showed how anomalous ENSO-related diabatic heating influences near-surface winds in the tropical Atlantic. This remote influence directly induces changes in the intensity of both the Atlantic Walker and Hadley circulations. The simulation of Mediterranean climate as influenced by some major modes of atmospheric variability appears to have improved in the recent years (see also Luterbacher et al., this book; Trigo et al., this book). In particular, the simulation of NAO responses to ENSO was improved. However, the nature of NAO prospects for predictability are limited to a few months and do not offer much field for predictability studies in the seasonal/interannual range. In this respect, a notable portion of the NAO predictability potential for future studies lies at scales longer than decades (Griffies and Bryan, 1997). Several reasons may account for the limited ability of the GCM to simulate the ENSO responses at mid-latitudes. Among those, note, for instance: ENSO transmission to mid-latitudes appears to operate through a complex teleconnection pattern that interacts with strong internal mid-latitude atmospheric dynamics. This transition further complicates its observational identification with techniques that need study of aggregates or ‘‘composite’’ events. This fact may also result in a serious limitation of its predictability potential. For instance, occasionally different events have been documented to have yielded different responses. The coarse resolution of global models over the Mediterranean region does not yet yield credible simulation scenarios. The nesting of regional models in global models is not yet developed enough for the Mediterranean sector, though together with downscaling techniques provides a promising area to investigate in the future. The Mediterranean sea is not adequately integrated in most model simulations. In addition, boundary responses coming from adjacent oceanic and terrestrial regions surrounding the Mediterranean area are not fully covered in regional experiments, yielding a poor representation of Mediterranean conditions. Processes of the previous four items may be responsible for some difference in ENSO sensitivity detected by observational and modelling studies. The latter usually yields weaker responses to ENSO out of the tropical regions
Climate Variability in the Mediterranean Region and the Tropics
159
(Rodo´, 2001). A deficient integration of transients and transitory couplings might account for a significant portion of the residual variability left, as proved by recent observational studies. Recent modelling studies show new areas for exploration in ENSO teleconnections, which might be of use in the future, in the search for increasing predictability of the Mediterranean climate. This is, for instance, the case with experiments seeking to simulate the atmospheric forcing in regions of the tropical North Atlantic during ENSO events (Lau and Nath, 2001). Another possibility is increasing horizontal resolution to obtain more reliable responses. This is the case for a study by Merkel and Latif (2002), illustrating that an increase of the horizontal resolution (from T42 to T106) causes significant changes in sea level pressure (SLP), temperature and precipitation over the Mediterranean as well as in the transient/stationary wave activity. A southward shift of the North Atlantic low pressure systems in the winter season during El Nin˜o was also noticed.
2.3. South Asian Monsoon Variability and the Mediterranean Climate The South Asia Monsoon (SAM) is a key factor influencing the climate of the eastern and central Mediterranean (Reddaway and Bigg, 1996; Rodwell and Hoskins, 1996; Ziv et al., 2004a,b). It causes high variability in SLP over Arabia and the Middle East with high pressures in winter and low pressures in summer. The adjustment to the SAM couples the falling pressure and land temperature over the Indian subcontinent/Asia Minor, with rising pressure and temperature over the Persian Gulf and Iraq. Another possible explanation for the different climatic behaviour of the eastern and western Mediterranean basins is derived from the gradual delay, of up to two weeks, of the onset of the monsoon in the 1980s, as compared with that in the early 1950s (Subbaramayya et al., 1990). This places the period of monsoon low pressure firmly in the summer months (JJA), whereas, previously, it was partly in May. On an average, this potentially lowers the summer pressure along with the temperature by shifting the monsoonal cloud cover, later in the season (Reddaway and Bigg, 1996). In accordance with Kripalani and Kulkarni (1999), this monsoonal delay could be attributed to the prolongation of the winter snow cover over Eurasia. They reported on a significant negative (positive) relationship between the wintertime snow depth over western Eurasia (eastern Eurasia and central Siberia) and subsequent Indian monsoon rainfall. This correlation structure is indicative of a mid-latitude longwave
160 Mediterranean Climate Variability
pattern with an anomalous ridge (trough) over Asia, during the winter prior to a strong (weak) monsoon. Rodwell and Hoskins (1996) showed that the Asian Summer Monsoon dominates not only Central Asia, but also the Eastern Mediterranean (EM). By using numerical simulations, they pointed at the linkage between the appearance of the semi-permanent subsidence structure over the EM and the onset of the Monsoon. The climatic regime and the dynamic factors governing the EM in the summer season, and their relationships with the Asian Monsoon, were analyzed by Ziv et al. (2004a), who found significant correlation on the interdiurnal timescale. They identified a circulation connecting the upward motion maximum over the Himalayas with the downward motion over the EM. Raicich et al. (2003) studied the relationship between the Asian and African Monsoon systems and found a high correlation between the intensity of each of them and the pressure distribution over the Mediterranean on the interannual timescale. The monsoon–desert mechanism presented by Rodwell and Hoskins (1996) may not be confined to the Asian monsoon alone. In a similar way, it could explain the relationship between the observed summertime strengthening of the oceanic sub-tropical anticyclones and the existence of western continental deserts and of ‘‘Mediterranean type’’ climate regions. They showed that the monsoon could force a remote descent to its west and northwest. The very dry summertime climate of the Mediterranean and the surrounding lands may be strongly related to this. They also showed that this descent is highly dependent on the latitude of the monsoon heating; a southward shift, for example, may lead to wetter weather, for southern Europe.
2.3.1. Mediterranean Climate and South Asian Rainfall The Indian summer Monsoon index has been recorded for almost 200 years, while records of the subsequent winter rain in Israel are relatively ‘‘younger’’; the longest record used is the one kept in Jerusalem, for the past 118 years. The overall correlation between these two indices was found to be only 0.3 (for the past 118 years). However, in 73 years (62%), the indices sign were the opposite. For extreme summer seasons, in which the index deviates by over 1.3 standard deviations, the correlation increases to 0.56 (Alpert et al., 2005). Similar results were found for other relatively long-record of rainfall stations in Israel. This illustrates the potential of the Indian Monsoon as a predictor for Israeli rainfall in the subsequent winter season. An important index of monsoon precipitation is the All-India Rainfall Index (AIR; Parthasarathy et al., 1995). It is an areal average of rainfall for 29 sub-divisions, which come from areally averaged district rainfalls. Rainfall
Climate Variability in the Mediterranean Region and the Tropics
161
amounts are totals for June, July, August and September (Parthasarathy et al., 1995). The AIR data are available online at: http://grads.iges.org/india/ allindia.html. Liu and Yanai (2001) found significant positive correlation between June–September AIR and JJAS tropospheric temperature from 1949 to 1998, over the entire Mediterranean and northern Africa, within pressure levels from 200 and 500 hPa levels. Similar results have been revealed for the southern and eastern Mediterranean in June and for the eastern Mediterranean in August (Liu and Yanai, 2001). While the role of the Tropical Atlantic Variability (TAV), ENSO and associated changes in SST over the tropical Pacific and Atlantic oceans have been widely investigated, the effect of the Indian Ocean on monsoon rainfall is not well understood. The existence of the Indian Ocean Dipole (IOD) mode was demonstrated by Saji et al. (1999) Webster et al. (1999) and Andersen (1999). A respective index was determined, though no statistical relationship between the index and the monsoon rains has been established. It is suggested that the variations in distribution and intensity of the EM rainfall, during the last decade, are associated with variations in the characteristics of the air mass over the Indian Ocean via its transport toward the EM. However, recent findings of idealized SST anomaly experiments by Hoerling et al. (2004) and Hurrell et al. (2004), indicate that SST variations have significantly controlled the North Atlantic circulation, related to the NAO, with the warming of the tropical Indian and western Pacific Ocean being of particular importance. When the winter regime over the entire Mediterranean is considered, the focus is given to the Rossby waves and other extra-tropical factors (such as the NAO) as the dominating features. However, some attention should be given to continental polar outbreaks associated with the Asian Winter Monsoon (e.g. Saaroni et al., 1996).
2.4. African Monsoon Impact on the Climate of the Mediterranean The climatic variables in the various parts of the Mediterranean are correlated with each other as well as with external circulations. For instance, the Mediterranean SLP oscillation (MO), i.e. the difference between its western and eastern parts, is correlated with precipitation. In winter, a fundamental role is played by the NAO index, whereas in summer, the regional Hadley cell was found to be correlated with climatic conditions over parts of the Mediterranean (see Trigo et al., this book). There is also some evidence for teleconnections with the South Asian Monsoon and with the Sahel precipitation. The correlation between the precipitation indices of these two systems and the MO is negative
162 Mediterranean Climate Variability
over the EM and positive over the western Mediterranean. The relevant governing mechanisms have been studied by several authors (see Baldi et al., 2002 for an extended bibliography), as well as the influence of the position and the strength of the Hadley cell (Dima and Wallace, 2003). Focusing on the summer season, Chen et al. (2002), showed evidence for strengthening of the tropical general circulation in the 1990s, and in particular the West Africa monsoon, reaching its northernmost extension in August, when the ITCZ, after the abrupt shift at the end of June and further slow northward migration, reaches its northernmost location (Sultan and Janicot, 2000, 2003). Important mechanisms, such as heat and moisture advection in North America and Asia and anomalously high values of the surface albedo in northern Africa, limit a further extension towards northern latitudes (Chou and Neelin, 2003; Rodwell and Hoskins, 1996, 2001). The two regimes, the dry and hot summers in the Mediterranean and the monsoon regime over West Africa, are highly correlated; interactions and feedback mechanisms between the two are not only possible, but also evident (Rowell, 2003; Baldi et al., 2002, 2003a,b). Ziv et al. (2004a), in their study of the summer regime, found a signature of the Hadley cell over eastern North Africa, connecting the EM with the African Monsoon. The relationship between them is manifested by a significant correlation between the ascent at 15 N–20 N latitudes and the descent at 30 N–40 N latitudes. The correlation between the EM subsidence and the Asian Monsoon was further validated through correlating the inter-diurnal variations of the vertical velocities of the two Monsoon systems, i.e. the Asian and the African Monsoons, yielding r ¼ 0.33, in spite of the 6000 km distance.
2.5. Tropical Cyclones’ Impact on the Mediterranean Climate Reale et al. (2001) showed that several cases of severe floods over the western Mediterranean could be traced back to hurricanes. Also, Hoskins and Berrisford (1988) related the severe 1987 storm in South England to hurricanes. Next, we review a first study (Krichak et al., 2004) showing the relationship between flooding in Israel and hurricanes (Fig. 49). Over the period from 3–5 December, 2001, there were heavy rains in northern Israel reaching 250 mm in some areas. The rains were associated with a relatively weak cyclone system approaching the area from the north-west. Atmospheric developments that produced the unusually intense rainfall and flash floods in Israel during 3–5 December 2001 were associated with upper-tropospheric jet stream activity. This activity was stimulated by the potential vorticity (PV) streamer conditions in the upper troposphere and by the intense intrusion of cold stratospheric air masses into the troposphere over the Mediterranean Sea area. Local topography and geography
Climate Variability in the Mediterranean Region and the Tropics
163
Figure 49: The back-trajectory from Mediterranean to Hurricane Olga 3 December–November 2001. of the EM region also played a role of an additional triggering factor in the process. The intense synoptic processes of December 2001 were initiated by the development of a tropical storm, which subsequently developed into hurricane Olga (from 25 to 29 November) accompanied by intense ascent motions in the tropical Atlantic. Convergence of huge amounts of atmospheric water vapour
164 Mediterranean Climate Variability
took place during the first stage of the hurricane development. Both the rise of large amounts of warm and moist tropical air and the subsequent release of latent heat caused an additional intensification of the hurricane. This process also induced development of an anticyclone to the north-east of Olga. The ascending moist air from Olga was later transported to Europe and finally to the Mediterranean region by the high intense clockwise atmospheric circulation in the process of Olga’s decline. This process led to the southward propagation of the polar jet and to the establishment of a situation characterized by the tropopause fold PV streamer with an extrusion of cold upper-tropospheric and stratospheric air over the south Alpine and the central Mediterranean areas. Formation and intensification of the EM cyclone of 3–5 December 2001 was additionally stimulated by the interaction of the polar and subtropical jets over the region (Krichak et al., 2004).
2.6. Tropical Intrusions into the Mediterranean Basin Rains in the Mediterranean basin take place mainly during winter, most of which is associated with Mediterranean baroclinic cyclones. Winter Mediterranean cyclones have their origin in the North Atlantic synoptic systems, in secondary lows formed when upper troughs interact with local orography, and/or with low level baroclinicity over the northern Mediterranean coast. However, processes originating from tropical regime are also significant in its eastern part (Krichak et al., 1997a,b; Krichak and Alpert, 1998; Dayan et al., 2001; Kahana et al., 2002; Ziv et al., 2004b) and along its western part, in north western Africa (Knippertz et al., 2003). The Red Sea Trough (RST) is one of the impressive manifestations of mid-latitude–tropical interactions in the EM especially during autumn and spring. The intensity and duration of the EM rain-spells highly depend on the interactions between the upper and lowertropospheric jets as well as their positioning and orientation. Specific jet characteristics stimulate development of meso-scale convective complexes and cyclogenesis. Due to turbulence associated with strong wind shear, tropopause folding may allow intrusions of the stratospheric air into the troposphere. It was recently shown that frequencies of RST intrusions into the EM, have nearly doubled since 1970 from about 50 d/y to about 100 d/y (Fig. 50) (Alpert et al., 2004a,b). Another type of rainstorms originating from the tropics is associated with ‘‘tropical plumes’’. This is a long cloud band that extends from the ITCZ down to 30 N–40 N latitude, accompanied by a pronounced trough in the Subtropical Jet to its west combined with a ridge to the east, while no common distinct system
Climate Variability in the Mediterranean Region and the Tropics
165
Figure 50: The Red Sea Trough frequencies as totals per hydrological year (August to July) and cumulative monthly contributions (October to April). at the surface or at the 500 hPa level, was found. Ziv (2001) found that prior to such type of a rainstorm the ‘‘tropical plume’’ is generated. It extends toward the subtropics, injects moisture of tropical origin that is captured by the Subtropical Jet, and if a pronounced trough develops there, extensive stratified cloudiness and widespread rains result. Zangvil and Isakson (1995) found in a rainstorm of the same type that the vertically integrated moisture convergence reached 1.8 mmh 1 over Israel, mostly above the 750 hPa level. Dayan and Abramski (1983) found an abnormal feature in the Subtropical Jet structure, i.e. a reversed position of its axis that leads to the formation of a large and humid warm air mass up to very high levels in the atmosphere above the Middle East.
2.7. Mediterranean Dust Transport from Sahara The role of atmospheric aerosols on the climate system is found to be most significant (IPCC, 2001). The dust radiative effect strongly depends on its vertical location. Daily model-based forecasts of 3D-dust fields could be used in order
166 Mediterranean Climate Variability
to determine the dust radiative effect in climate models, because of the large gaps in observations of dust vertical profiles (Alpert et al., 2004c). The averaged dust vertical distribution, based on the 3-year database of 48-hour dust forecasts, shows significant differences between the Atlantic and the Mediterranean dust transport. As a whole, the Mediterranean dust is found to be within a wider range of altitudes, penetrating high into the troposphere (Fig. 51). Supporting evidence for this characteristic feature of the Mediterranean dust transport was obtained from the analysis of lidar dust profiles over Rome (Italy), collected in the 3-year period 2001–2003 during the high dust activity season from March to June (Kishcha et al., 2005). Based on the data set of dustaffected lidar profiles (206), Fig. 52 presents histograms of the main parameters of these dust layers. In particular, the bottom boundary was found to range from 0.5 to 5 km, with the mean value BT ¼ 1.6 0.8 km; the top boundary ranges from 2.4 to 8 km, with mean value TP ¼ 5.1 1.1 km, and the thickness of dust layers ranges from 0.4 to 7.5 km, with mean value TH ¼ 3.6 1.5 km. Hence, on an average, dust over Rome is distant from the surface and penetrates high into the troposphere. Moreover, as shown in Fig. 52, the Gaussian fitting curves suit the histograms of lidar-derived data. In seasons other than March–June, some indication of the mean vertical distribution of dust over Rome can be found in Gobbi et al. (2004), based on lidar data collected in the year 2001.
(10
Figure 51: Latitudinal cross-sections of averaged dust concentrations 7 kg/m3) for the months of April, zonal averaged within the longitudinal zone 30 E–40 E. Adapted from Alpert et al. (2004c).
Climate Variability in the Mediterranean Region and the Tropics
167
A
B
C
Figure 52: Statistical distributions of lidar-derived parameters of the dust layer over Rome from March to June based on the data set of dust-affected lidar profiles (206) between 2001 and 2003: bottom (A) and top (B) heights (km), and thickness, km (C). Fitting curves of the Gaussian distribution are shown by dotted lines. From Kishcha et al. (2005).
168 Mediterranean Climate Variability
The lidar vertical profiles collected in the presence of dust over Rome were also used in order to validate the Tel Aviv University (TAU) dust model. A quantitative comparison of model vertical profiles against lidar soundings was made and the model was found good in about 70% of the cases (Kishcha et al., 2005). Saharan dust is generally transported over the Mediterranean by southerly winds generated by cyclones (Alpert and Ziv, 1989; Bergametti et al., 1989; Alpert et al., 1990; Moulin et al., 1998). In particular, Alpert and Ziv (1989) found that spring and early summer are the most favourable periods for the development of Saharan lows (also called Sharav cyclones) south of the Atlas Mountains. Usually, such cyclones move eastward and cross Egypt, Israel and the eastern Mediterranean basin. As shown by Bergametti et al. (1989) and Moulin et al. (1998), dust outbreaks to the western and central parts of the Mediterranean are linked with two depression centres: Saharan lows and a high over Libya. The high over Libya prevents Saharan lows from following an eastward direction. This synoptic situation, having a peak in spring and in early summer, induces strong south and southwestern winds between the two systems and is characterized by dust intrusions from North Africa to the Mediterranean basin. Moreover, complex wind fields associated with frontal zones under those atmospheric conditions could be one of the causal factors for dust over the Mediterranean being within a wide range of altitudes, penetrating high into the troposphere, as mentioned above. The mean synoptic situation associated with dust outbreaks from Sahara into the central Mediterranean was examined on a daily basis for the month of July from 1979 to 1992 (Barkan et al., 2005). It was found that the strength and position of two essential features of the circulation patterns, such as the trough emanating southward from the Iceland low and the eastern cell of the subtropical high, are the governing factors in making suitable flows for the Saharan dust transportation toward Central Europe. The typical composite pattern of wind in the case of five days of great quantity of dust in the atmosphere above Italy between 5 and 9 July 1988 is shown in Fig. 53. A deep low over Ireland with a strong trough emanating from it southward and splitting the subtropical high into two separate cells is apparent. The eastern high pressure centre is located over Sicily. Between the Irish low and the Sicilian high, a strong southwesterly flow transports dust from Mauritania across the western Mediterranean to central Europe.
2.8. Conclusions and Outlook The aforementioned evidence of tropical teleconnections to the Mediterranean climate suggests further analysis in order to test these relationships by using
Climate Variability in the Mediterranean Region and the Tropics
169
Figure 53: Average wind flow of the dusty period 5–9 July 1988 at 700 hPa. This period corresponds to the Saharan dust transportation toward Central Europe. Adapted from Barkan et al. (2005).
appropriate modelling and statistical methodologies. The factor separation method (Stein and Alpert, 1993; Alpert et al., 1995) may be useful for distinguishing among contributions of several factors and also of their synergetic effects in producing weather patterns over the Mediterranean. Thus, the modelling approach with a well-defined methodology is necessary for a clear and simple mechanistic understanding of the different teleconnections discussed here.
170 Mediterranean Climate Variability
2.8.1. Future Research on ENSO Impact on Mediterranean Climate Investigate the role of mid-latitude ocean in responding to the atmospheric forcing which have a tropical origin (Lau and Nath, 2001) and its effect on the Mediterranean climate. Improve resolution and accuracy of observational studies with the use of a denser, homogeneous set of instrumental records. Implement new statistical techniques capable of address local phenomena. These are needed to address ENSO and other tropical influences in the Mediterranean climate. As an example, the new Scale-Dependent Correlation (SDC) technique (Rodo´, 2001; Rodo´ et al., 2002; Rodrı´ guez-Arias and Rodo´, 2004) may be useful. Analyse and devise modelling experiments which can cope with a complex response to ENSO, also through the alteration of internal modes of variability at mid-latitudes (e.g. NAO, EATL-WRUS, etc.). Improve the nesting of regional climate models, increase their horizontal resolution and refine model simulations for a more realistic representation of the Mediterranean climate. Explore the different scenarios of the future ENSO frequency and intensity changes, in response to climate change (e.g. Timmermann et al., 1999). Assess their relation to the Mediterranean climate variability and extremes.
2.8.2. Future Research on South Asian Monsoon Variability and the Mediterranean Climate To study teleconnections of the South Asian Monsoon with the eastern Mediterranean for different time scales, i.e. interannual, seasonal and decadal time scales. Attempt to evaluate the range of influence of the Asian Monsoon over the entire Mediterranean basin. To study long-term trends of various variables, as Saaroni et al. (2003) and Ziv et al. (2005) performed for summer temperature, and relate them to long-term trends in the South Asian Monsoon features along the entire year. To study the detailed structure of summer circulations over the eastern Mediterranean region prior to and during extreme episodes in which the EM undergoes heat waves or exceptional rain events. To incorporate data about the South Asian Monsoon into the seasonal prediction scheme for the Israel winter rainfall. To validate the suggested linkages between the Indian Ocean processes and the eastern Mediterranean climate.
Climate Variability in the Mediterranean Region and the Tropics
171
To develop a climatologic basis for continental polar outbreaks events over the Mediterranean. This includes both synoptic and statistical detailed analyses. To assess the statistical relationship between the variations in the EM rainfall amount, distribution and intensity, on the one hand, and the long-range variations of the characteristics of the air mass transport associated with the Indian Ocean Dipole, on the other hand.
2.8.3. Future Research on African Monsoon Impacts on the Climate of the Mediterranean To study teleconnections between the summer climate in the Mediterranean and the African Monsoon by using numerical simulations. The major tools could be the NCEP–NCAR and ECMWF reanalyses, historical time series of atmospheric parameters in southern Europe (Luterbacher et al., this book), Regional numerical models, scenarios for future climate produced by global climate models, like the ones from the Canadian Centre for Climate Modelling and Analysis (CCCma), and also gridded precipitation data provided by the Global Precipitation Climatology Project. To perform numerical simulations with the Regional Model on different time-space scales for the domain including Europe, the Mediterranean Basin and the northern part of the African continent north to the Gulf of Guinea. The effects of SST variability in the Gulf of Guinea on the climate variability in the Mediterranean should be assessed by using an approach similar to that presented by Vizy and Cook (2001, 2002). In turn, the influence of the Mediterranean SST on climate variability in the North African region should be studied. To perform time-slice experiments for the future climate evolution by using the regional model, according to different available scenarios. Since the phenomena are embedded in the large scale circulation and in particular in the Hadley cell circulation, therefore a mathematical model of the evolution of the Hadley cell should be elaborated. To study the linkage between the Mediterranean climate, CLIVAR VACS (Variability of the African Climate System) and AMMA (African Monsoon Multidisciplinary Activities).
2.8.4. Future Research on Tropical Intrusions into the Mediterranean Basin To define general mechanisms of tropical intrusions of the Red Sea trough and the tropical plume into the EM.
172 Mediterranean Climate Variability
To find out the role of the Red Sea trough and the tropical plume in the general atmospheric circulation over the Mediterranean. In particular, to find out their role in the transport of moisture and angular momentum. To study physical reasons and mechanisms of the recent increase in tropical intrusions into the Mediterranean.
Acknowledgements The study at Tel Aviv University was supported by the GLOWA-Jordan River BMBF-MOS project and the Israeli Science Foundation (ISF), grant no. 828/02.
References Allan, R., Lindesay, J., & Parker, D. (Eds). (1996). El Nino-Southern Oscillation and Climate Variability. CSIRO Publishing, Collingwood, Victoria, Australia, 416 p. Alpert, P., & Ziv, B. (1989). The Sharav cyclone: observations and some theoretical considerations. J. Geophys. Res., 94, 18495–18514. Alpert, P., Neeman, B. U., & Shay-El, Y. (1990). Intermonthly variability of cyclone tracks in the Mediterranean. J. Climate, 3, 1474–1478. Alpert, P., Stein, U., & Tsidulko, M. (1995). Role of sea fluxes and topography in Eastern Mediterranean cyclogenesis. The Global Atmosphere-Ocean System, 3, 55–79. Alpert, P., Ben-Gai, T., Baharad, A., Benjamini, Y., Yekutieli, D., Colacino, M., Diodato, L., Ramis, C., Homar, V., Romero, R., Michaelides, S., & Manes, A. (2002). The paradoxical increase of Mediterranean extreme daily rainfall in spite of decrease in total values. Geophys. Res. Lett., 29(11), 31-1–31-4, (June issue). Alpert, P., Ilani, R., da-Silva, A., Rudack, A., & Mandel, M. (2005). Seasonal prediction for Israel winter precipitation based on northern hemispheric EOF. MERCHAVIM special issue for Prof. A. Bitan, in press. Alpert, P., Osetinsky, I., Ziv, B., & Shafir, H. (2004a). Semi-objective classification for daily synoptic systems: application to the EM climate change. Int. J. Climatol., 24, 1001–1011. Alpert, P., Osetinsky, I., Ziv, B., & Shafir, H. (2004b). A new seasons definition based on the classified daily synoptic systems: an example for the EM. Int. J. Climatol., 24, 1013–1021. Alpert, P., Kishcha, P., Shtivelman, A., Krichak, S. O., & Joseph, J. H. (2004c). Vertical distribution of Saharan dust based on 2.5-year model predictions. Atmos. Res., 70, 109–130. Andersen, D. (1999). Extremes in the Indian Ocean. Nature, 401, 337–339. Baldi, M., Crisci, A., Dalu, G. A., Maracchi, G., Meneguzzo, F. & Pasqui, M. (2002). Mediterranean climate and its connections to regional and global processes. Proceedings of the First Italian IGBP Conference Mediterraneo e Italia nel Cambiamento Globale: un ponte fra scienza e societa’ Paestum, Salerno, 14–16 November 2002.
Climate Variability in the Mediterranean Region and the Tropics
173
Baldi, M., Capecchi, V., Crisci, A., Dalu, G. A., Maracchi, G., Meneguzzo, F. & Pasqui, M. (2003a). Mediterranean summer climate and its relationship to regional and global processes. Proceedings of the Sixth European Conference on Applications of Meteorology, Rome, 15–19 September 2003. Baldi, M., Capecchi, V., Crisci, A., Dalu, G.A., Maracchi, G., Meneguzzo, F., & Pasqui, M. (2003b). Numerical analysis of the teleconnection of the West Africa monsoon with the mediterranean summer climate. Environmental Fluid Mechanics, submitted. Barkan, J., Alpert, P., Kutiel, H., & Kishcha, P. (2005). The synoptics of dust transportation days from Africa toward Italy and Central Europe. Journal of Geophysical Research – Atmosphere, 110, DO 7208, doi: 10.1029/2004JD005222. Bergametti, G., Dutot, A. L., Buat-Menard, P., Losno, R., & Remoudaki, E. (1989). Seasonal variability of the elemental composition of atmospheric aerosol particles over the northwestern Mediterranean. Tellus, 41B, 353–361. Chen, J., Carlson, B. E., & Del Genio, A. D. (2002). Evidence for strengthening of the tropical general circulation in the 1990s. Science, 295, 838–841. Chou, C., & Neelin, J. D. (2003). Mechanisms limiting the northward extent of the northern summer monsoons over North America, Asia, and Africa. J. Climate, 16, 406–425. Compo, G. P., Sardeshmukh, P. D., & Penland, C. (2001). Changes of subseasonal variability associated with El Nin˜o. J. Climate, 14, 3356–3374. Curtis, S., & Hastenrath, S. (1995). Forcing of anomalous SST evolution in the tropical Atlantic during Pacific warm events. J. Geophys. Res., 100, 15835–15847. Dayan, U., & Abramski, R. (1983). Heavy rain in the Middle East related to unusual Jet Stream properties. Bulletin of the American Meteorological Society, 64, 1138–1140. Dayan, U., Ziv, B., Margalit, A., Morin, E., & Sharon, D. (2001). A severe autumn storm over the Middle-East: synoptic and mesoscale convection analysis. Theor. Appl. Clim., 69, 103–122. Diaz, H. F., Hoerling, M. P., & Eischeid, J. K. (2001). ENSO variability, teleconnections and climate change. Int. J. Climatol., 21, 1845–1862. Dima, I. M., & Wallace, J. M. (2003). On the seasonality of the Hadley Cell. J. Atmos. Sci., 60, 1522–1527. Enfield, D. B., & Mayer, D. A. (1997). Tropical Atlantic SST variability and its relation to El Nin˜o Southern Oscillation. J. Geophys. Res., 102, 929–945. Fraedrich, K. (1994). ENSO impact on Europe? – a review. Tellus, 46A, 541–552. Fraedrich, K., & Mu¨ller, K. (1992). Climate anomalies in Europe associated with ENSO extremes. Int. J. Climatol., 12, 25–31. Giorgi, F. (2002). Variability and trends of sub-continental scale surface climate in the twentieth century. Part I: observations. Clim. Dyn., 18, 675–691, doi: 10.1007/ s00382–001-0204-x. Gobbi, G. P., Barnaba, F., & Ammannato, L. (2004). The vertical distribution of aerosols, Saharan dust and clouds at Rome (Italy) in the year 2001. Atmos. Chem. Phys., 3, 2161–2172. Griffies, S. M., & Bryan, K. (1997). Predictability of North Atlantic climate variability. Science, 275, 181–184. Gualdi, S., Navarra, A., Guilyardi, E., & Delecluse, P. (2003). The SINTEX coupled GCM. The tropical Indo-Pacific region. Ann. Geophys., 46, 1–26.
174 Mediterranean Climate Variability
Halpert, M. S., & Ropelewski, C. F. (1992). Surface temperature patterns associated with the SO. J. Climate, 5, 577–593. Hoerling, M. P., Hurrell, J. W., Xu, T., Bates, G. T., & Phillips, A. (2004). Twentieth century North Atlantic climate change. Part II: understanding the effect of Indian Ocean warming. Clim. Dyn., 23, 391–405, doi: 10.1007/s00382-004-0433-x. Hoskins, B. J., & Berrisford, P. (1988). A potential vorticity perspective of the storm of 15–16 October 1987. Weather, 42, 122–129. Huang, J., Higuchi, K., & Shabbar, A. (1998). The relationship between the North Atlantic Oscillation and the El Nin˜o-Southern Oscillation. Geophys. Res. Lett., 25, 2707–2710. Hurrell, J. W., Hoerling, M. P., Phillips, A. S., & Xu, T. (2004). Twentieth century North Atlantic climate change. Part I: assessing determinism. Clim. Dyn., 23, 371–389, doi: 10.1007/s00382-004–0432-y. IPCC, Intergovernmental Panel on Climate Change (2001). The Scientific basis (Contribution of WGI to the 3rd Assessment Report of the IPCC), Cambridge University Press, Cambridge. Kadiog˘lu, M., Tulunay, Y., & Borhan, Y. (1999). Variability of Turkish precipitation compared to El Nin˜o events. Geophys. Res. Lett., 26, 1597–1600. Kahana, R., B. Ziv, Enzel, Y., & Dayan, U. (2002). Synoptic climatology of major floods in the Negev Desert, Israel. Int. J. Clim., 22, 822–867. Kiladis, G. N., & Dı´ az, H. F. (1989). Global climatic anomalies associated with extremes in the Southern Oscillation. J. Climate, 2, 1069–1090. Kishcha, P., Barnaba, F., Gobbi, G. P., Alpert, P., Shtivelman, A., Krichak, S. O., & Joseph, J. H. (2005). Vertical distribution of Saharan dust over Rome (Italy): comparison between 3-year model predictions and lidar soundings. J. Geophys. Res., 110, D6208. doi:10.1029/2004JD005480. Klein, S. A., Soden, B. J., & Lau, N.-C. (1999). Remote sea surface temperature variations during ENSO: evidence for a tropical atmospheric bridge. J. Climate, 12, 917–932. Knippertz, P., Fink, A. H., Reiner, A., & Speth, P. (2003). Three late summer/early autumn cases of tropical-extratropical interactions causing precipitation in Northwest Africa. Mon. Wea. Rev., 131, 116–135. Krichak, S. O., Alpert, P., & Dayan, M. (2004). The role of atmospheric processes associated with hurricane Olga in the December 2001 floods in Israel. J. Hydrometeorology, 5(6), 1259–1270. Krichak, S. O., Alpert, P., & Krishnamurti, T. N. (1997a). Interaction of topography and tropospheric flow – a possible generator for the Red Sea trough? Meteorol. Atmosph. Phys., 63(3–4), 149–158. Krichak, S. O., Alpert, P., & Krishnamurti, T. N. (1997b). Red Sea trough/cyclone development – numerical investigation. Meteorol. Atmosph. Phys., 63(3–4), 159–170. Krichak, S. O., & Alpert, P. (1998). Role of large scale moist dynamics in November 1–5 1994 hazardous Mediterranean Weather, 103(D16), 19453–19468. Kripalani, R. H., & Kulkarni, A. (1999). Climatology and variability of historical Soviet snow depth data: some new perspectives in snow – Indian monsoon teleconnections. Clim. Dyn., 15, 475–489. Laita, M., & Grimalt, M. (1997). Vorticity and pressure anomalies in the western Mediterranean during El Nin˜o/Southern Oscillation extremes. Int. J. Climatol., 17, 475–482.
Climate Variability in the Mediterranean Region and the Tropics
175
Lanzante, J. R. (1996). Lag relationships involving tropical sea surface temperatures. J. Climate, 9, 2568–2578. Lau, N.-C., & Nath, M. J. (1996). The role of the ‘atmospheric bridge’ in linking tropical Pacific ENSO events to extratropical SST anomalies. J. Climate, 9, 2036–2057. Lau, N.-C., & Nath, M. J. (2001). Impact of ENSO on SST variability in the North Pacific and North Atlantic: seasonal dependence and role of extratropical sea-air coupling. J. Climate, 14, 2846–2866. Liu, X., & Yanai, M. (2001). Relationship between the Indian monsoon rainfall and the tropospheric temperature over the Eurasian continent. Quart. J. Roy. Met. Soc., 127, 909–938. Lloyd-Hughes, B., & Saunders, M. A. (2002). Seasonal prediction of European spring precipitation from El Nino-Southern Oscillation and local sea-surface temperatures. Int. J. Climatol., 22, 1–14. Loon, H., & Madden, R. A. (1981). The Southern Oscillation. Part I: global associations with pressure and temperature in Northern winter. Mon. Wea. Rev., 109, 1150–1168. Mariotti, A., Zeng, N., & Lau, K. M. (2002). Euro-Mediterranean rainfall and ENSO – a seasonally varying relationship. Geophys. Res. Lett., 29, doi 10.1029/2001GL014248. Mariotti, A., Ballabrera-Poy, J., & Zeng, N. (2005). Tropical influence on Euro-Asian autumn rainfall variability. Clim. Dyn., 24(5), 511–521, 10.1007/s00382-004–0498-6. Mason, S. J., & Goddard, L. (2001). Probabilistic precipitation anomalies associated with ENSO. Bulletin of the Americ. Meteorol. Soc., 82(4), 619–638. Merkel, U., & Latif, M. (2002). A high resolution AGCM study of the El Nino impact on the North Atlantic/European sector. Geophys. Res. Lett., 29(9), 1291. Mestas-Nunez, A. M., & Enfield, D. B. (2001). Eastern Equatorial Pacific SST variability: ENSO and non-ENSO components and their climatic associations. J. Climate, 14, 391–402. Moron, V., & Ward, M. N. (1998). ENSO teleconnections with climate variability in the European and African sectors. Weather, 53, 287–295. Moulin, C., Lambert, C., Dayan, U., Masson, V., Ramonet, M., Bousquet, P., Legrand, M., Balkanski, Y., Guelle, W., Marticorena, B., Bergametti, G., & Dulac, F. (1998). Satellite climatology of African dust transport in the Mediterranean atmosphere. J. Geophys. Res., 103, 13137–13144. Nobre, P., & Shukla, J. (1996). Variations of SST, wind stress, and rainfall over the tropical Atlantic and South America. J. Climate, 9, 2464–2479. Oldenborgh, G. J., Burgers, G., & Klein-Tank, A. (2000). On the El Nin˜o teleconnection to spring precipitation in Europe. Int. J. Climatol., 20, 565–574. Parthasarathy, B., Munot, A. A., & Kothawale, D. R. (1995). Monthly and seasonal rainfall series for all-India homogeneous regions and meteorological subdivisions: 1871–1994. Contributions from Indian Institute of Tropical Meteorology. Research Report RR-065, Aug. 1995, Pune, India. Pozo-Va´zquez, D., Esteban-Parra, M. J., Rodrigo, F. S., & Castro-Diez, Y. (2001). The association between ENSO and winter atmospheric circulation and temperature in the North Atlantic region. J. Climate, 14, 3408–3420. Price, C., Stone, L., Rajagopalan, B., & Alpert, P. (1998). A possible link between El Nin˜o and precipitation in Israel. Geophys. Res. Lett., 25, 3963–3966. Raicich, F., Pinardi, N., & Navarra, A. (2003). Telleconections between Indian Monsoon and Sahel rainfall and the Mediterranean. Int. J. Climatol., 23, 173–186.
176 Mediterranean Climate Variability
Reale, O., Feudale, L., & Turato, B. (2001). Evaporative moisture sources during a sequence of floods in the Mediterranean region. Geophys. Res. Lett., 28, 2085–2088. Reddaway, J. M., & Bigg, G. R. (1996). Climatic change over the Mediterranean and links to the more general atmospheric circulation. Int. J. Climatol., 16, 651–661. Rocha, A. (1999). Low-frequency variability of seasonal rainfall over the Iberian Peninsula and ENSO. Int. J. Climatol., 19, 889–901. Rodo´, X. (2001). Reversal of three global atmospheric fields linking changes in SST anomalies in the Pacific, Atlantic and Indian oceans at tropical latitudes and midlatitudes. Clim. Dyn., 18, 203–217. Rodo´, X., & Comin, F. A. (2000). Links between large-scale anomalies, rainfall and wine quality in the Iberian Peninsula during the last three decades. Global Change Biology, 6(3), 267–273. Rodo´, X., Baert, E., & Comı´ n, F. A. (1997). Variations in seasonal rainfall in Southern Europe during the present century: relationships with the NAO and the ENSO. Clim. Dyn., 13, 275–284. Rodo´, X., Pascual, M., Fuchs, G., & Faruque, A. S. G. (2002). ENSO and cholera: a nonstationary link related to climate change? PNAS, 99, 12901–12906. Rodrı´ guez-Arias, M. A., & Rodo´, X. (2004). A primer on the study of transitory dynamics in ecological series using the scale-dependent correlation analysis. Oecologia, 138(4), 485–504. Rodriguez-Puebla, C., Encinas, A. H., Nieto, S., & Garmendia, J. (1998). Spatial and temporal patterns of annual precipitation variability over the Iberian Peninsula. Int. J. Climatol., 18, 299–316. Rodwell, M. J., & Hoskins, B. J. (1996). Monsoons and the dynamic of deserts. Quar. J. Roy. Meteorol. Soc., 122, 1385–1404. Rodwell, M. J., & Hoskins, B. J. (2001). Subtropical anticyclones and summer monsoons. J. Climate, 14, 3192–3211. Ropelewski, C. F., & Halpert, M. S. (1987). Global and regional precipitation patterns associated with El Nin˜o-Southern Oscillation. Mon. Wea. Rev., 115, 1606–1626. Rowell, D. P. (2003). The impact of Mediterranean SSTs on the Sahelian rainfall season. J. Climate, 16, 849–862. Ruiz-Barradas, A., Carton, J. A., & Nigam, S. (2003). Role of the atmosphere in climate variability of the tropical Atlantic. J. Climate, 16, 2052–2065. Saaroni, H., Bitan, A., Alpert, P., & Ziv, B. (1996). Continental outbreaks into the Levant and Eastern Mediterranean. Int. J. Clim., 16, 1–17. Saaroni, H., Ziv, B., Edelson, J., & Alpert, P. (2003). Long-term variations in summer temperatures over the Eastern Mediterranean. Geo. Res. Lett., 30(18), 1946, doi:10.1029/2003GL017742. Saji, N. H., Goswami, B. N., Vinayachandran, P. N., & Yamagata, T. (1999). A dipole mode in the tropical Indian Ocean. Nature, 401, 360–363. Saravanan, R., & Chang, P. (2000). Interaction between tropical Atlantic variability and ENSO. J. Climate, 2177–2194. Stein, U., & Alpert, P. (1993). Factor separation in numerical simulations. J. Atmos. Sci., 50, 2107–2115. Subbaramayya, I., Vivekanandababu, S., & Naidu, C. V. (1990). Variations in the onset of the Indian south-west monsoon and summer circulation anomalies. Meteorol. Magaz., 119, 61–67.
Climate Variability in the Mediterranean Region and the Tropics
177
Sultan, B., & Janicot, S. (2000). Abrupt shift of the ITCZ over West Africa. Geophys. Res. Letter, 27, 3353–3356. Sultan, B., & Janicot, S. (2003). The West African monsoon dynamics. Part II the preonset and the onset of the summer monsoon. Journal of Climate, 16(21), 3407–3427. Sutton, R. T., Jewson, S. P., & Rowell, D. P. (2000). The elements of climate variability in the tropical Atlantic region. J. Climate, 13, 3261–3284. Timmermann, A., Oberhuber, J., Bacher, A., Esch, M., Latif, M., & Roeckner, E. (1999). Increase El Nin˜o frequency in a climate model forced by future greenhouse warming. Nature, 398, 694–697. Trenberth, K. E., & Hoar, T. J. (1997). El Nin˜o and climate change. Geophys. Res. Let., 24, 3057–3060. Trenberth, K. E., Branstator, W., Karoly, D., Kumar, A., Lau, N., & Ropelewski, C. (1998). Progress during TOGA in understanding and modeling global teleconnections associated with tropical sea surface temperatures. J. Geophy. Res., 103, 14291–14324. Tu¨rkes, M. (1998). Influence of geopotential heights, cyclone frequency and southern oscillation on rainfall variations in Turkey. Int. J. Climatol., 18, 649–680. Vizy, E. K., & Cook, K. H. (2001). Mechanisms by which Gulf of Guinea and Eastern North Atlantic sea surface temperature anomalies can influence African rainfall. J. Climate, 11, 3167–3191. Vizy, E. K., & Cook, K. H. (2002). Development and application of a mesoscale climate model for the tropics: influence of sea surface temperature anomalies on the West African monsoon. J. Geophys. Res., 107(D3), 10.1029/2001JD000686. Wallace, J. M., & Gutzler, D. (1981). Teleconnections in the geopotential height fields during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784–812. Webster, P., Moore, J. A. M., Loschnigg, J. P., & Leben, R. R. (1999). Coupled ocean atmosphere dynamics in the Indian Ocean during 1997–98. Nature, 401, 356–360. Wuethrich, B. (1995). El Nino goes critical. New Scientist, 4, (Feb), 32–35. Xoplaki, E. (2002). Climate variability in the Mediterranean. Ph.D. thesis, University of Bern, Switzerland, http://sinus.unibe.ch/klimet/docs/phd_xoplaki.pdf. Yakir, D., Lev-Yadun, S., & Zangvil, A. (1996). El Nino and tree growth near Jerusalem over the last 20 years. Global Change Biology, 2, 101–105. Zangvil, A., & Isakson, A. (1995). Structure of the water vapor field associated with an early spring rainstorm over the Eastern Mediterranean. Isr. J. Earth Sci., 44, 159–168. Ziv, B. (2001). A subtropical rainstorm associated with a tropical plume over Africa and the Middle-East. Theor. Appl. Clim., 69(1/2), 91–102. Ziv, B., Saaroni, H., & Alpert, P. (2004a). The factors governing the summer regime of the Eastern Mediterranean. Int. J. Clim., 24, 1859–1871. Ziv, B., Dayan, U., & Sharon, D. (2004b). A mid-winter, tropical extreme floodproducing storm in southern Israel: synoptic scale analysis. Meteorology and Atmospheric Physics, doi, 10.1007/s00703-003–0054-7. Ziv, B., Saaroni, H., Baharad, A., Yekutieli, D., & Alpert, P. (2005). The indications for aggravation in summer heat conditions over the Mediterranean Basin. Geophysical Research Letters, 32(12), L12706, 10.1029/2005GL022796.
This page intentionally left blank
Chapter 3
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability Ricardo Trigo,1 Elena Xoplaki,2 Eduardo Zorita,3 Ju¨rg Luterbacher,2 Simon O. Krichak,4 Pinhas Alpert,4 Jucundus Jacobeit,5 Jon Sa´enz,6 Jesu´s Ferna´ndez,3 Fidel Gonza´lez-Rouco,7 Ricardo Garcia-Herrera,7 Xavier Rodo,8 Michele Brunetti,9 Teresa Nanni,9 Maurizio Maugeri,10 Murat Tu¨rkes° ,11 Luis Gimeno,12 Pedro Ribera,13 Manola Brunet,14 Isabel F. Trigo,15 Michel Crepon16 and Annarita Mariotti17 1
CGUL at University of Lisbon and Universidade Luso´fona, Portugal (
[email protected]) 2 Institute of Geography and NCCR Climate, University of Bern, Bern, Switzerland (
[email protected],
[email protected]) 3 GKSS Research Centre, Germany (
[email protected],
[email protected]) 4 Tel Aviv University, Israel (
[email protected],
[email protected]) 5 Institute of Geography, University of Augsburg, Germany (
[email protected]) 6 Department of Applied Physics II, University of the Basque Country, Spain (
[email protected]) 7 Dto Fı´sica de la Tierra II, Universidad Complutense, Madrid, Spain (
[email protected],
[email protected]) 8 ICREA and Climate Research Laboratory, PCB-University of Barcelona, Spain (
[email protected]) 9 Institute of Atmospheric Sciences and Climate, National Research Council, Bologna, Italy (
[email protected],
[email protected]) 10 Istituto di Fisica Generale Applicata, University of Milan, Milan, Italy (
[email protected]) 11 Department of Geography, Faculty of Sciences and Arts, C¸anakkale Onsekiz Mart University, C¸anakkale, Turkey (
[email protected]) 12 Facultad de Ciencias de Ourense, Universidad de Vigo, Ourense, Spain (
[email protected]) 13 Departamento de Ciencias Ambientales, Universidad Pablo de Olavide de Sevilla, Ctra Utrera, Km 1, Spain (
[email protected])
180 Mediterranean Climate Variability
14
Climate Change Research Group, University Rovira i Virgili, Tarragona, Spain (
[email protected]) 15 Instituto de Meteorologia and CGUL at University of Lisbon (
[email protected]) 16 LODYC BC 100, Universite P.-et-M.-Curie, Paris, France (
[email protected]) 17 Earth System Science Interdisciplinary Center, USA and ENEA, Italy (
[email protected])
3.1. Introduction The Mediterranean climate is under the influence of both tropical and midlatitude climate dynamics, being directly affected by continental and maritime air masses with significant origin differences (Barry and Chorley, 2003). In this region, most of the precipitation occurs from October to March (Xoplaki, 2002). The peak of the winter season occurs between December and February, when the mid-latitude cyclone belt has usually reached its southernmost position (e.g. HMSO, 1962). However, spring and autumn also contribute to a significant amount of precipitation (Fig. 54). Being located at the southern limit of the North Atlantic storm tracks, the Mediterranean region is particularly sensitive to interannual shifts in the trajectories of mid-latitude cyclones that can lead to remarkable anomalies of precipitation and, to a lesser extent, of temperature. Given the seasonal characteristics of the Atlantic storm-tracks, this is particularly true in winter when the influence of mid-latitude variability is at its greatest. In the transition seasons, and especially in summer, this influence needs to be considered along with other factors including that of tropical climate. Storm-track variability impacts primarily the western Mediterranean, but it has also a signature clearly detected in the eastern Mediterranean as well. The complex orography that characterizes most regions surrounding the Mediterranean basin can modulate and even distort climate anomaly patterns that otherwise would be geographically much more homogenous. For instance, the interplay between orography and thermal contrast between advected Atlantic air masses and Mediterranean temperatures has a huge impact on the development of Mediterranean storms (Trigo et al., 2002a), which can produce violent precipitation extremes at the end of the summer season. Observational studies indicate significant climate trends on different time scales in the Atlantic–European area, including the larger Mediterranean area. The physical processes responsible for these trends and changes seem to be hemispheric to global (such as external forcings and changes in the large-scale
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
181
Figure 54: Seasonal distribution of precipitation over the entire Mediterranean Basin according to the monthly database from the Global Historical Climatology Network (GHCN). The stations from GHCN were randomly subsampled to evenly cover the area and the common period 1948–1990 was used (Adapted from Ferna´ndez et al., 2003). atmospheric circulation) as well as local/regional (such as changes in land surface and use). It is one of the main challenges to understand the recent trends and changes over the Mediterranean region, both in space and time. Furthermore, it is necessary to study these relatively recent trends within a larger temporal framework. Such analyses was performed in Chapter 1 of this book. Why is it so relevant to study the physical mechanisms responsible for variability and trends of both temperature and precipitation over the Mediterranean? There is evidence that major changes on the strength of some of these circulation modes have already made their impact on the living conditions of many people around the Mediterranean basin (see Chapter 1) and that future changes on these patterns will probably produce significant changes in the regional climate in the future (see Chapter 8). Furthermore, it must be stressed that there has always been a considerably large number of people living in this area with a strong dependence on regional agriculture and availability of water resources. Agriculture still constitutes a major economic activity in the Mediterranean region, particularly for southern Mediterranean countries, precisely those countries that are most affected by the variability of water availability. Critical situations of water shortages and extended droughts are mostly due to high values of seasonal and year-to-year variability of
182 Mediterranean Climate Variability
precipitation. The impact of temperature and precipitation on most crop yields is mostly dependent on changes in the seasonal cycle of these parameters, rather than on fluctuation of their annual average value (Ro¨tter and van de Geijn, 1999). In this temperate climate, the most important variable is related to the presence or lack of water. Lack of water in winter and spring will be reflected in the crop yield. However, too much water in winter can be harmful by drowning the seeds and retarding root development (Xoplaki et al., 2001). The variability of precipitation plays a crucial role in the management of regional agriculture, in environment, in water resources and ecosystems as well as social development and behaviour (Xoplaki, 2002).
3.2. Mid-Latitude Modes of Atmospheric Variability and their Impact It is now widely accepted that most large-scale modes of atmospheric circulation in the Northern Hemisphere have been described previously in literature (Wallace and Gutzler, 1981; Barnston and Livezey, 1987). It should be stressed that the relevance of these modes is seasonally dependent, i.e. they have a signature only during part of the year (Barnston and Livezey, 1987). Different approaches have been developed over the last decade to assess the impact of the most relevant modes on the Mediterranean climate, mostly in terms of precipitation and temperature fields. Generally speaking, these studies can be clustered within two different approaches (Yarnal, 1993): (a) Studies based on atmospheric circulation indices independently from surface climate parameters. These include the pioneering work on blocking episodes by Rex (1950a,b, 1951) and, more recently on the North Atlantic Oscillation pattern (Hurrell, 1995). At shorter spatio-temporal scales, there are various regional classification schemes such as the Lamb Weather Types for the UK (Lamb, 1972) and the Grosswetterlagen catalogues (Hess and Brezowski, 1977) for central Europe. (b) Methods that incorporate both atmospheric circulation and surface climatic fields, often based on eigenvalue techniques (e.g. CCA, SVD). Some of these works have focused on the Mediterranean basin (e.g. Corte-Real et al., 1995; Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2000, 2003a,b, 2004; Valero et al., 2004). Despite their different methodologies, these studies tend to agree that the most important mid-latitude modes for the Mediterranean climate at the monthly time-scale are: (a) the North Atlantic Oscillation (NAO), (b) the Eastern Atlantic
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
183
(EA) pattern and Eastern Atlantic/Western Russia pattern (EA/WR; EU2 of Barnston and Livezey, 1987), and (c) the Scandinavian pattern (SCAND; EU1 of Barnston and Livezey, 1987). Naturally we will focus our analysis on the impact of these modes that affect most significantly the Mediterranean Basin. It is worth noticing that the impact of the most important circulation patterns on the Mediterranean climate in historical times was addressed in Chapter 1 of this book. Traditionally, studies on large-scale patterns and climate trends have been carried out by slightly different research communities, with the former being mostly developed by dynamical meteorologists and the latter mainly by climatologists and geographers. This framework has changed considerably in the last two decades, but it is still reflected in the way this chapter was organized, with some issues analysed from more than one viewpoint. Section 3.2 describes the most important large-scale modes that have been recognized to impact significantly the climate of the Mediterranean region. Sections 3.3 and 3.4 analyse the influence of these circulation modes in the variability throughout the twentieth century of the temperature and precipitation fields. Section 3.5 addresses again the impact of these patterns to account for the observed trends of climatic variables within the Mediterranean basin. Despite our best efforts, we acknowledge that with this sequential approach, some repetitions of issues have become inevitable.
3.2.1. The North Atlantic Oscillation Pattern The North Atlantic Oscillation has been identified for more than 70 years as one of the major patterns of atmospheric variability in the Northern Hemisphere (Walker, 1924). However, it has only become the subject of a wider interest in recent years (e.g. van Loon and Rogers, 1978; Rogers, 1984; Barnston and Livezey, 1987; Lamb and Peppler, 1987; Hurrell, 1995; Hurrell and van Loon, 1997; Wanner et al., 2001). This mode is associated with the strength of the meridional pressure gradient along the North Atlantic sector (Fig. 55, top). Some authors consider the NAO mode as a regional manifestation of the hemispheric Arctic Oscillation (Thompson and Wallace, 1998). This discussion is out of the scope of this chapter, nevertheless, it highlights how climate anomalies in the Mediterranean region need to be considered in the larger context of global climate variability – and that we should look at the larger picture when evaluating the impact of the NAO on the Mediterranean climate. Moreover, Xoplaki (2002) has shown that the impact of the NAO and the AO patterns in Mediterranean winter precipitation and temperature fields is quite similar.
184 Mediterranean Climate Variability
Figure 55: Top: Difference in SLP (hPa, solid contours) and NCEP precipitation rate (mm/day, colour) between winter months with a NAO index >1 and months with an NAO index < 1 (period 1958–1997). Precipitation rate differences are represented only if significant at the 5% level. Bottom: As in top but with high resolution precipitation field (mm/day) of New et al. (2000) (represented only if significant at the 5% level) (Adapted from Trigo et al., 2004a).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
185
Since the pioneering work by Lamb and Peppler (1987), most works for the Mediterranean area have been focused on the impact of the NAO during the winter season (December to March) when its impact is greatest, particularly for precipitation (Rodrı´ guez-Fonseca and Castro, 2002). This control exerted by NAO on the precipitation field is related to corresponding changes in the associated activity of North-Atlantic storm tracks that affect most of western Europe (Osborn et al., 1999; Ulbrich et al., 1999; Goodess and Jones, 2002; Trigo et al., 2002b) and the Eastern Mediterranean such as in Turkey (Tu¨rkes° and Erlat, 2003, 2005). Using high and (low) NAO index composites, Trigo et al. (2002b, 2004a) have shown anomaly fields of climate variables and their associated physical mechanisms for the entire Europe (Figs. 55 and 56). While
Figure 56: Top: Difference in maximum temperature ( C) between winter months with an NAO index >1 and months with an NAO index < 1 (period 1958–1997). Differences are represented only if significant at the 5% level. Bottom: As in top but with minimum temperature. Data from NCEP/NCAR reanalyses.
186 Mediterranean Climate Variability
the impact of the NAO on precipitation shows a decreasing influence in the eastern and southern sector of the Mediterranean basin (Mariotti et al., 2002b), it can still impact significantly on precipitation variability and water resources over Turkey (Cullen and de Menocal, 2000; Struglia et al., 2004; Tu¨rkes° and Erlat, 2003, 2005). The influence of the NAO on the variability of the whole Mediterranean sea–fresh water cycle is linked to the precipitation anomalies since no significant correlation is found with evaporation (Mariotti et al., 2002b). This NAO-precipitation control is associated with the steering of storm-track paths over the entire North-Atlantic sector, but also influences cyclogenesis in the Mediterranean (Trigo et al., 2000). This issue is further developed in Chapter 6 in this book. It has been shown that the NAO does not play a relevant role in terms of western Mediterranean winter temperature variability (Sa´enz et al., 2001b; PozoVazquez et al., 2001a; Castro-Dı´ ez et al., 2002) and a minor, but discernible, one for the eastern Mediterranean temperature field (Cullen and de Menocal, 2000; Ben Gai et al., 2001; Xoplaki, 2002). Generally, the influence of the positive (negative) winter NAO on the Mediterranean is warmer (cooler) conditions over the northern part and cooler (warmer) over the southern part (Hurrell, 1995; Trigo et al., 2002b; Xoplaki, 2002). Interestingly, the impact of the NAO in daily extreme temperatures is unequal, with large asymmetries between minimum and maximum temperatures (Fig. 56), and more significantly, between positive and negative phases of NAO (Trigo et al., 2002b). The differences in maximum and minimum temperatures between months of high and low NAO index are shown in Fig. 56. The amplitude of the differences over Europe is larger for minimum than for maximum temperatures. However, the spatial extension of statistically significant differences for Iberia and southern Europe in general is larger for maximum than for minimum temperature. It is worth noting that maximum temperature values are usually recorded during daylight while minimum temperature values are usually observed towards the end of the night. Thus, during daytime, enhanced solar short wave radiation is capable of partially offsetting the advection of cold polar air to yield small maximum temperature anomalies, while during the night, the strong clear sky emission of long wave radiation further cools the lower troposphere (Trigo et al., 2002b). The large-scale mean temperature anomalies can be mostly explained by heat transport by the corresponding anomalous mean atmospheric flow, modified and partially offset by the heat transported by transient eddies. However, there is a third process: the modulation by anomalous cloud cover (associated with anomalous atmospheric circulation) of the radiative transfer of heat to and from the Earth’s surface. These radiative and cloud cover influences modulate the response to NAO mainly in terms of generating different day and night-time temperature anomalies (Trigo et al., 2002b, 2004b).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
187
3.2.2. The Eastern Atlantic and Eastern Atlantic/Western Russia Patterns Several authors have shown that modes other than NAO play an important role in shaping the European precipitation variability, including sectors of the Mediterranean Basin (Zorita et al., 1992; von Storch et al., 1993; Qian et al., 2000; Quadrelli et al., 2001; Krichak et al., 2002; Xoplaki, 2002). However the NAO pattern is sufficiently well established and has been derived through different methods for every month of the year. The remaining modes of atmospheric circulation over Europe present a less clear picture, because they are of a more regional nature, their index may not be so unambiguous and finally, because some of these modes have a signature only during part of the year (Barnston and Livezey, 1987). In particular, the Eastern Atlantic (EA) pattern depends crucially on the procedure used to derive it. Still, the kind of variability associated with this pattern seems important and physically real, as it is also detected in studies using alternative techniques, like cluster analysis (Kimoto and Ghil, 1993). According to Wallace and Gutzler (1981), the EA corresponds to an index, defined in terms of the geopotential height anomalies of the 500 hPa surface at three different points. Other definitions of the same index and ‘‘similar’’ patterns have been provided through the years, including the indices from rotated EOF analysis (Barnston and Livezey, 1987). These authors (Barnston and Livezey, 1987) have identified two patterns, the Eastern Atlantic and the Eastern Atlantic/Western Russia (EA/WRUS). The East Atlantic/ Western Russia pattern is one of the two prominent patterns that affect Eurasia during most of the year presenting its east Atlantic anomaly centre located further east than the EA (over western Europe) and an opposite centre located north of the Caspian region. This pattern is prominent in all months except June–August, and has been referred to as the Eurasia-2 pattern by Barnston and Livezey (1987). Correlation coefficients obtained between monthly time series of the EAWG (defined by Wallace and Gutzler, 1981) and EABL or EA/WRUS (as defined by Barnston and Livezey, 1987) are statistically significant but relatively low. For instance, the correlation value between monthly winter (D,J,F) of EAWG and EABL is 0.54 (significant at the 95% level) and between EAWG and EA/WRUS is 0.46. Therefore, results of impact studies that use slightly different indices are bound to obtain different correlation. Here, we decided to adopt a broad perspective on this issue, making an effort to include results obtained by authors that have used both EA as well as the EA/WRUS indices. The subscripts from EAWG and EABL are dropped in the following, because different studies adopt one of the definitions for the EA index they use. Recent works have shown that the EA and EA/WRUS patterns represent a significant contribution for the precipitation over northern Iberia (Sa´enz et al., 2001a) as well as parts of the eastern Mediterranean areas (Quadrelli et al., 2001;
188 Mediterranean Climate Variability
Xoplaki et al., 2000, 2004; Krichak et al., 2002). Using station data, Xoplaki (2002) shows the spatial correlation between the EA/WRUS and winter (NDJF) precipitation in the Mediterranean basin (Fig. 57, top). Significant positive correlation is visible between EA/WRUS and winter precipitation over northeastern Africa, the Near East, eastern Turkey and the Black Sea region. Significant negative correlations are found generally north of 40 N with a maximum over France. Anomalous positive pressure over northwestern Europe and the central and western Mediterranean area lead to subsidence and stability conditions and reduced precipitation over continental Europe and the northern coast of the Mediterranean (Xoplaki, 2002). The advection of humid and warm air by EA/WRUS to certain Mediterranean regions is associated with
Figure 57: Spatial Spearman correlation between the patterns EA/WRUS and SCAND patterns and winter (NDJF) Mediterranean station precipitation for the period 1950–1999. Correlations |r|0.14 indicate significance at the 95% level, |r|0.18 at the 99% level and |r| 0.23 at the 99.9% level (n ¼ 4 50 ¼ 200 months), respectively (Adapted from Xoplaki, 2002).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
189
the southward shifts of storm tracks from western Europe towards the Mediterranean. This effect, combined with the local cyclogenesis, is the main physical mechanism responsible for increased precipitation values in the Mediterranean (Xoplaki et al., 2004). In the case of winter (DJF) temperature variability, it has been shown that for the western Mediterranean region, the fraction of variance explained by the EA mode is higher than the one explained by the NAO pattern. This result is linked to the sensible heat fluxes by the mean circulation anomalies driven by the EA pattern (Sa´enz et al., 2001b). A recent work has shown that the EA also plays a remarkable role in shaping daily temperature variability over most of north-eastern Spain throughout the year, particularly over coastal areas (Sigro´, 2004), with the correlation coefficient reaching its maximum value in February (0.70). However, results as high as these are not observed for the eastern Mediterranean (Hasanean, 2004).
3.2.3. The Scandinavian and Blocking Patterns The Scandinavian (SCAND) pattern consists of a primary circulation centre, which spans Scandinavia and large portions of the Arctic Ocean north of Siberia (Xoplaki, 2002). Two additional weaker centres with opposite sign to the Scandinavian centre are located over western Europe and over the Mongolia/ western China sector. The SCAND pattern is a prominent mode of low frequency variability in all months except June and July, and has been previously referred to as the Eurasia-1 pattern by Barnston and Livezey (1987). The SCAND pattern is associated with important precipitation anomalies in both western and eastern Mediterranean regions (Corte-Real et al., 1995; Wibig, 1999; Xoplaki, 2002; Quadrelli et al., 2001). Using station data, Xoplaki (2002) obtained the spatial pattern of correlation between the SCAND mode and winter (NDJF) precipitation in the Mediterranean basin (Fig. 57, bottom). A strong positive pressure anomaly centered over Scandinavia and western Russia and negative anomalies over the Iberian Peninsula cause anomalous easterly to southeasterly airflow over the eastern basin and anomalous southwesterlies to southerlies over the central basin. The combined effect of these air masses connected with the relatively warm Mediterranean Sea leads to distinct cyclogenesis connected with high precipitation amounts over Italy, along the eastern Adriatic coast and the southern part of the Alps. Interestingly, the SCAND circulation mode reveals a spatial pattern similar to the European blocking pattern, usually described in studies using sub-monthly scales (e.g. Tibaldi et al., 1997). Therefore, we believe that it is appropriate to describe the impacts of the well-established blocking pattern (Rex, 1950b, 1951) simultaneously with those associated with the SCAND mode. In fact,
190 Mediterranean Climate Variability
a recent study has quantified the impact of the SCAND pattern winter blocking variability over the European sector (Barriopedro et al., 2005). The frequency of blocking events over Europe presents a marked seasonal cycle with higher values in winter and spring (Tibaldi et al., 1997; Trigo et al., 2004b). Blocking episodes are known to produce significant impacts on both the precipitation and temperature fields of the Mediterranean Region (Trigo et al., 2004b). Figure 58 (top) shows differences between the mean 500 hPa geopotential height composites for winter blocking and non-blocking episodes, and the corresponding difference for the 850 hPa temperature composites (represented only if significant at the 1% level). The corresponding differences for precipitation rate (represented only if significant at the 5% level) are shown in Fig. 58 (bottom). Blocking episodes usually last between 5 and 20 days, but their fingerprint is sufficiently intense to be noticed at the monthly scale
Figure 58: Top: Differences between the mean 500 hPa geopotential height (gpm, contour) composites for winter blocking and non-blocking episodes, and the corresponding difference for the 850 hPa ( C, colour) temperature composites (represented only if significant at the 1% level). Bottom: Corresponding differences of precipitation rate (mm/day) are represented (only if locally significant at the 5% level). Data from NCEP/NCAR reanalyses (Adapted from Trigo et al., 2004b).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
191
(Quadrelli et al., 2001). The most important feature corresponds to the intensification of the meridional component of the mid-troposphere circulation that is perfectly visible during blocked situations, up and downstream of the British Isles. This configuration is usually associated with the split of the jet stream in two distinct branches, a feature that is widely accepted as a trademark of European blocking episodes (Rex, 1950a). The impact of these events for the low tropospheric temperature field at 850 hPa extends from Iberia to the Black Sea with the most intense values being observed over the Balkans (Fig. 58, top). On the other hand, the significant impact of these blocking episodes on the Mediterranean precipitation is confined to the western sector (Fig. 58, bottom).
3.2.4. Other Modes A regional manifestation of the NAO is given by the Mediterranean Oscillation (Conte et al., 1989) implying opposite pressure (especially at upper levels) and surface climate conditions between the western to central and the southeastern Mediterranean basin. Thus, the most important canonical correlation pattern between the large-scale circulation and Mediterranean precipitation in winter being significantly correlated (r ¼ 0.72) with the NAO during the October–March period is strongly related with various indices of the Mediterranean Oscillation (Du¨nkeloh and Jacobeit, 2003). A similar result arises even for the case of a nonseasonal whole-year analysis at monthly scales (Corte-Real et al., 1995). Therefore, such a high correlation coefficient between NAO and the Mediterranean Oscillation seems to imply that these two phenomena are not independent. Another dominant circulation mode (present at 1000 and 500 hPa) coupled with Mediterranean climate variability is the Mediterranean Meridional Circulation (MMC) pattern consisting of two opposite anomaly centres west of the Bay of Biscay and in the central Mediterranean implying preferred meridional flows around this area (Du¨nkeloh and Jacobeit, 2003). It shows some relation with the hemispheric-scale EA pattern at the 500 hPa level (NOAA-CPC, 2005a) and often recurs in dynamical studies linked to the Mediterranean area, e.g. as second mode of the non-seasonal analysis by Corte-Real et al. (1995) or as first mode of the Greek winter rainfall analysis by Xoplaki et al. (2000). Another mode of low-frequency variability affecting the Mediterranean area is the East Atlantic Jet pattern (EA-JET, see NOAA-CPC, 2005a). It is among the ten leading teleconnection patterns from April to August (NOAA-CPC, 2005b) revealing a N–S dipole of anomaly centres with one centre over the northeastern North Atlantic and western Scandinavia, the other centre over Northwest Africa and large parts of the Mediterranean region. Thus, its positive mode reflects intensified mid-latitude westerlies, whereas its negative mode represents
192 Mediterranean Climate Variability
particular blocking configurations. During spring, there is a moderate correlation (r ¼ 0.44) with a canonical correlation pattern (CCP) whose rainfall part explains some 16% of Mediterranean precipitation variability, however, during summer the EA-JET pattern is strongly related (r ¼ 0.66) to the most important CCP including nearly 30% of explained summer precipitation variability in the northern and western Mediterranean regions (Du¨nkeloh and Jacobeit, 2003). Thus, blocking configurations of an EA-JET type in its negative mode are especially important for above-average summer rainfall in these areas. Touchan et al. (2005) recently have also found a significant positive correlation between the EA-Jet and May–August precipitation over the southeastern Mediterranean area for the 1948–2000 period. However, they also report on instationarities in those relationships using reconstructed precipitation several centuries back in time. A last mode, the Polar Eurasian pattern (NOAA-CPC, 2005a), is particularly important during the winter season (NOAA-CPC, 2005b). It reflects variations in the strength of the circumpolar vortex and affects the Mediterranean region by its European anomaly centre which extends up to the northern and western parts of the Mediterranean area. Increased anticyclonicity in these regions represents an enhanced polar vortex and vice versa. This pattern has been found being moderate negatively correlated ( 0.42) with the second CCP of wet season Mediterranean precipitation connected with below normal precipitation over the western Mediterranean and above normal precipitation over the eastern part of the basin (Xoplaki et al., 2004).
3.3. Temperature Variability It is widely accepted that the frequency of large-scale circulation patterns has a major impact on monthly/seasonal surface climate characteristics. In fact this link was explored in Chapter 1 of this book to evaluate the evolution of Mediterranean winter climate since the early sixteenth century. However, until the mid-twentieth century, virtually all large-scale circulation data is referent to sea level pressure, limiting the extent of those historical circulation-climate studies (except for Schmutz et al., 2000, Luterbacher et al., 2002 and Bro¨nnimann and Luterbacher, 2004, who provide mid and upper tropospheric fields further back in time). Nowadays, the generalized use of multi-variable and multi-level reanalyses datasets has allowed the use of more appropriate variables (e.g. 300 hPa or 500 hPa geopotential height) to establish those relationships. Xoplaki et al. (2003b) using a multi-component CCA in the EOF space, investigated the relationships between the large-scale atmospheric circulation and the Mediterranean summer (June to September) temperature. The authors
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
193
show that 56% of the summer Mediterranean temperature variability during the second half of the twentieth century can be explained by making use of the first two canonical modes. The most important of the canonical modes (Fig. 59) reveals a dipole configuration in the North Atlantic. In the positive phase of the dipole, a deep centre with positive 300 hPa geopotential height anomalies is located over central Europe; an area of negative anomalies presents higher values south of Iceland and surrounding northward, the positive anomalies extends up to the western Ural mountains and the northern Caspian Sea. SSTs and surface temperatures reveal higher values in the northwestern part of the Mediterranean under the high pressure region. This dynamic configuration strengthens the zonal flow over northern Europe and easterly–northeasterly flow over the Mediterranean. The increased stability leads to clear sky conditions and maximum insolation in the area. The CCA mode described above agrees well with the second mode in the study of Xoplaki et al. (2003a) on summer temperatures over Greece. Thus, it seems plausible to extend the reasoning in their work to the entire Mediterranean area. From this perspective, the variability of summer temperatures in the Mediterranean would be well described with a parsimonious conceptual model invoking two modes, the ‘‘high-index’’ type and the ‘‘low-index’’ type, which picture transitions from the zonal to the meridional flow.
3.4. Precipitation Variability Most Mediterranean countries experience frequent drought episodes, which may cause water shortages and disrupt agricultural and industrial activities, such as hydroelectric power generation. Ko¨ppen’s (1936) definition of Mediterranean climate is, in simple terms, one in which winter rainfall is more than three times the summer rainfall. Summer in many regions located in the southern Mediterranean coast is characterized by lack of rain, long periods of drought and high temperatures that lead to a marked summer aridity. The strong summer–winter rainfall contrast that characterizes the Mediterranean climate is associated with pronounced seasonal cycles in most climatic variables. During a typical year, rain occurs most frequently during the winter half-year over most of the land area surrounding the Mediterranean, mainly in the southern and eastern parts. In addition, the winter half-year precipitation accounts for between 30% (western and northern Mediterranean lands) and 80% (easternsoutheastern parts) of the annual total amounts (Xoplaki et al., 2004). In general, there exists a clear deficit of water during the summer half-year, when only sparse storms and convective systems produce rainfall (Trigo et al., 1999).
194 Mediterranean Climate Variability
Figure 59: Canonical spatial patterns of the first multi-component CCA. The canonical correlation patterns reflect the typical strength of the signal, with 300 hPa, 700–1000 hPa, SST, air-temperature anomalies in C (adapted from Xoplaki et al., 2003b).
195
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
50 N
50 N
0.0
0.4
0.4 0.2
0.2 0.40.6
0.6
40 N
0.4
0.8
0.4
0.4 0.2 0.0
0.6 0.2
0.4 0.4
0.0
-0.2
0.0 -0.2
0.4
-0.2
-0.4
0
10 E
0.2
0.2
0.6
0.2
0.2 0.0
0.2
0.4 0.2
40 N
.6
30 N 10 W
0.2
0.2
0.2
0.6
0.6
20 E
30 E
40
50 N
0.4 0.2
30 N 10 W
0
10 E
20 E
30 E
40
50 N 0.2 0.4
0.2
0.0
0.4 0.2
0.0 0.2
0.6
0.6
0.0
0.4
0
10 E
20 E
50 N
0.2
40 N
0.6 0.4
0.6 0.4 0.2 0.0
-0.2
30 N 10 W
0.2
0.4
0.6
-0.2
40 N
0.2
0.0 0.2
0.0
30 E
0.4
0.2
0.2
40
30 N 10 W
0
10 E
20 E
30 E
40
20 E
30 E
40
50 N
0.2
0.4 0.6
0.0 -0.2 -0.4
0.4
0.2 0.4
0.4
40 N
0.0
0.6
0.2
-0.2
0.2
0.0
-0.4
-0.2
-0.2
40 N
-0.2 -0.2
-0.4
0.2
0.0
-0.2
-0.2
30 N 10 W
0
10 E
20 E
30 E
40
30 N 10 W
0
10 E
Figure 60: Left: From top to bottom: The 3 leading EOFs from the New et al. (2000) precipitation field expressed as correlation of the corresponding PC with the precipitation series at each grid point. Right: From top to bottom: The variance fraction of the precipitation series explained by the 3 leading PCs. (Adapted from Ferna´ndez et al., 2003).
Therefore, it is necessary to use appropriate statistical tools to assess the spatio-temporal variability of the precipitation field throughout the yearly cycle, but particularly during the winter season. Figure 60 shows the approach developed for the Mediterranean basin by Ferna´ndez et al. (2003) in which PCA was applied to the linearly detrended monthly precipitation over the area given by the high resolution (0.5 0.5 ) gridded data from New et al. (2000) for the period 1948–1996. The 3 leading EOFs are represented as the correlation of the corresponding PC with the precipitation series at each gridpoint. These three patterns explain around 50% of the total detrended field. It is worth mentioning that the first EOF (22% of variance) shows the typical precipitation fingerprint
196 Mediterranean Climate Variability
of the NAO, controlling the rainfall variability in western Mediterranean sector (see Fig. 55). The second EOF (16% of variance) shows high loading values over the eastern Mediterranean sector while the third EOF (11% of variance) shows a north–south dipole with particularly high loadings over France. The spatial pattern of the third EOF is related with the patterns of precipitation impact of the EA or the EA/WRUS already mentioned (Fig. 57, left). The average flux of atmospheric water vapour through the boundary is shown in Fig. 61. Since the outflow is selected as positive flux, the negative values along the western boundary represent the inflow of humidity coming from the Atlantic Ocean. An average inflow of 494 Pg mo 1 crosses this boundary while the main outflow from the basin takes place through the eastern boundary (Ferna´ndez et al., 2003). The precipitation variability is closely related to the structure of the vertically integrated moisture transport fluxes, inside the domain and at the borders. As an example, the principal components of precipitation are regressed over the moisture transport at each NCEP/NCAR Reanalysis grid point over the area in Fig. 61B–D. The leading model (Fig. 61B) shows for positive phases an intensification of the transports of moisture arriving to the area from the Atlantic through the lateral western boundary. The second model A
C
B
D
Figure 61: (A) Average moisture flux through the boundaries according to NCEP data (in Pg mo 1). (B–D) Regression of the precipitation PCs (1st–3rd PC, respectively) over the vertically integrated moisture transport derived from NCEP/NCAR Reanalysis in kg m 1 s 1. (Adapted from Ferna´ndez et al., 2003).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
197
(Fig. 61C) shows an important contribution of moisture to the atmosphere from the western Mediterranean sub-basin and its transport to the eastern areas. Finally, the third EOF (Fig. 61D) is closely linked to the barrier effect of orography (mainly over the Alps) over the vertically integrated moisture transports. Xoplaki et al. (2004) used a multi-component Canonical Correlation Analysis (CCA) in the Empirical Orthogonal Function (EOF) space to identify the most important circulation patterns, at the sea level as well as at mid- and upper atmospheric levels, associated with the wet season (October to May; 1949–1999) Mediterranean station precipitation. Standard CCA technique relates one largescale pattern with one regional precipitation pattern (e.g. Corte-Real et al., 1995). However, Xoplaki et al. (2004) have shown that a combination of large-scale fields of predictors can achieve better results than the use of a single predictor. Four large-scale circulation modes accounted for 30% of the overall precipitation variability. It should be stressed that the CCA method maximizes the explained variance that links patterns of precipitation and large scale fields previously considered. Of course, there is a part of the variability linked to local factors and not related to these large-scale fields and, thus, the variance explained by these patterns (constrained to be related to the pre-defined large scale circulation patterns) is lower than the one associated with the first 3 EOFs by Fe´rnandez et al. (2003). The first canonical pair (Fig. 62) is connected with the NAO and
Figure 62: Canonical spatial patterns of the first CCA between the wet season Mediterranean station precipitation (predictand) and 300 hPa, 500 hPa and SLP (predictors); wet season precipitation anomalies in mm (Adapted from Xoplaki et al., 2004).
198 Mediterranean Climate Variability
the EA/WRUS (correlates at 0.66 and 0.50, respectively). These patterns are connected with above normal precipitation over the western, central and northern Mediterranean area and drier conditions over the remaining region. These results are quite well in agreement with Du¨nkeloh and Jacobeit (2003) for a similar analysis using gridded precipitation data over the Mediterranean coast. A regional study focused on the eastern Mediterranean supports that stronger westerlies over the eastern North Atlantic and the rising 500 hPa height (and the sea level pressure) over continental Europe during the last few decades were connected with enhanced atmospheric stabilization and anomalous advection of cold dry air from northerly directions. This led to the winter dryness over the eastern Mediterranean (Xoplaki et al., 2000).
3.5. Trends A comprehensive analysis of Mediterranean climate trends over the last 500 years has been presented in Chapter 1 of this book. Here, we focus on trends observed over shorter time scales (decadal) during the twentieth century. After describing the most important characteristics of temperature and precipitation trends for different parts of the Mediterranean region, we describe the most important physical mechanisms associated with them and driven by the large-scale atmospheric patterns described in the previous sections.
3.5.1. Temperature Trends Giorgi (2002) analysed the surface air temperature variability and trends over the larger Mediterranean land-area for the twentieth century based on gridded data of New et al. (2000). He found a significant warming trend of 0.75 C, mostly from contributions during the early and late decades of the century. Slightly stronger warming was observed for winter and summer. Spring reveals important trends over the last half century, mainly in the northern part of the Mediterranean while Autumn presents a warming trend in the western basin and a slight decrease in the eastern sector. The structure of climate series can differ considerably across regions showing variability at a range of scales in response to changes in the direct radiative forcing and variations in internal modes of the climate system (New et al., 2001; Hansen et al., 2001; Giorgi, 2002). Therefore, it is of no surprise, that based on the same data as Giorgi, Jacobeit (2000) found a distinct summer warming between 1969 and 1998 (a short period), being more distinct in the western than in the eastern part; seasonal cooling trends exist in some eastern areas, mainly in spring, in some cases also in winter. Figure 63 (left) presents the linear trends of summer
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
Figure 63: Right: Linear trends of winter (NDJF) station precipitation (mm/50 year). Left: Summer (JJAS) station air temperatures ( C/50 year) for the period 1950–1999. Stations with a significant trend (90% confidence level, based on the Mann–Kendall test) are encircled (from Xoplaki, 2002).
199
200 Mediterranean Climate Variability
station air temperatures ( C/50 year) for the period 1950–1999 based on the results of Xoplaki et al. (2002). It also shows the stations which experienced a significant trend. A clear east–west differentiation in Mediterranean summer air-temperature trends is visible. Negative values over the Balkans and eastern basin can be observed, however many are not significant. In the other areas, there is a significant warming trend of up to 3 C/50 year. However, the warming in these regions did not occur in a steady or monotonic fashion. Over most of western Mediterranean for instance, it has been mainly registered in two phases: the early years of twentieth century up to the mid-century warm phase and from the earliest 1970s onwards (Fig. 64). These rising episodes in temperatures were particularly well depicted over the entire Iberian Peninsula (e.g. Brunet et al., 2001; Galan et al., 2001) and over Italy (Brunetti et al., 2006). A glance at the summer air-temperature trends for the period 1900–1949 reveals that warming, though less extreme as in 1950–1999, was experienced in the western basin (not shown). A cooling trend over 1900–1949 was only prevalent over Libya and Egypt. The trend of winter temperature over 1900–1949 indicates a general cooling in the central basin but a warming in the east and west (not shown). For the 1950–1999 period, except for the eastern part, there was warming experienced (not shown). Xoplaki (2002) found a significant cooling trend of Mediterranean winter Sea Surface Temperatures (SSTs) east of 20 E over the period 1950–1999, while the western basin SST experienced a positive trend. Xoplaki et al. (2003b) showed that the 300 hPa geopotential height, 700– 1000 hPa thickness and Mediterranean SST large-scale fields account for more than 50% of the Mediterranean summer temperature variability over the period 1950–1999. The most important summer warming pattern is associated with
Figure 64: The curve corresponds to standardized values of the spatial average of Mediterranean summer temperatures for the period 1850–1999. All time series are outputs of a 10-year centred moving average filter (Adapted from Xoplaki et al., 2003b).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
201
blocking conditions, subsidence and stability. This mode is responsible for the 0.4 C (0.5 C) warming during the period 1950–1999 (1900–1999). The spatial average for summer (June to September) Mediterranean station temperatures (Fig. 64) shows high values during the 1860s, comparable to those during the 1950s and 1990s. For the period 1850 to 1999, a trend of 0.018 C/decade is found (significant at the 95% level). For the period 1900 to 1999, a change of 0.05 C/ decade is found. This fact points to an increase in temperature of about 0.27 C in the 1850–1999 period, and of 0.5 C in the twentieth century (Xoplaki et al., 2003b). For the eastern part of the Mediterranean, no significant linear trend in the averaged summer months (June to September) and entire summer mean air temperature could be detected (Xoplaki et al., 2003a). The most remarkable features concerning trends are on decadal time scales: the cooling trend at the beginning of the 1960s and the warming at the end of the 1980s (Fig. 65).
A
B
Figure 65: (A) Spatially weighted mean of summer (JJAS) air temperature anomalies from Greece and western Turkey from 1950 to 1999. Circles: 10 coolest summers. Squares: 10 warmest summers. Solid symbols correspond to the extreme summers, in which at least three single months are characterized by ‘‘extreme’’ air temperature conditions. (B) Spatially weighted mean of summer (JJAS) air temperature anomalies (Thessaloniki, Larissa, Athens and Patra) from 1901 to 1999 (From Xoplaki et al., 2003a).
202 Mediterranean Climate Variability
Tu¨rkes° et al. (2002) found general increasing trends in annual, winter and spring mean temperatures particularly over the southern regions of Turkey, for the period 1929–1999. On the other hand, these authors found decreasing trends for summer and autumn mean temperatures over the inner continental and northern regions. In general, summer night-time warming, temperature rates were found to be larger than corresponding night-time rates for spring and autumn. Furthermore, during spring and summer night-time warming, temperature rates were generally stronger than daytime temperature rates. Recently, Tu¨rkes° and Su¨mer (2004) showed that the diurnal temperature ranges significantly decreased for most urban stations of Turkey throughout the year, but less explicitly in winter.
3.5.2. Precipitation Trends It is a well-known fact that precipitation throughout the Mediterranean basin is highly concentrated in time between late autumn and spring (i.e. between October and April). Therefore, the relatively vast amount of literature on the subject is bound to reflect this imbalance, with very few studies focusing the drier half of the year. Recent studies revealed that the twentieth century was characterized by significant precipitation trends at different time and space scales (e.g. Folland et al., 2001; New et al., 2001). Giorgi (2002) found negative (positive) winter precipitation trends over the eastern (western) Mediterranean land-area for the twentieth century. The two other important seasons for precipitation present a different picture, with negative trends concentrated over Iberia and central Mediterranean areas (Scho¨nwise et al., 1993). This is confirmed with more regionalized studies, such as the one undertaken by Gonza´lez-Rouco et al. (2001) for the Iberian Peninsula, showing positive seasonal trend for this time interval in winter and negative trends in spring and autumn. In Italy, the significant negative trend observed in total annual precipitation amount is mainly due to the spring season, even if negative but not significant trends were observed also for winter, summer and autumn (Brunetti et al., 2006). Studies on precipitation trends must be analysed carefully as they crucially depend on the length of time-series analysed. Using the same data as Giorgi, Jacobeit (2000) showed for the last three decades some rainfall increases in autumn (western Iberia and southern Turkey), but dominating decreases in winter and spring. For the period 1951–2000, opposite to the prevailing decreasing trend, there is some increase in winter precipitation from southern Israel to northern Libya in accordance with increased positive modes of the Mediterranean Oscillation (Jacobeit et al., 2004). Nevertheless, the prevailing trends for the second half of the twentieth century are negative in winter and spring, although areas
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
203
with significance are relatively restricted, depending on the month (Norrant and Dougue´droit, 2005). A glance at the Mediterranean regional winter precipitation trends reveals a more detailed picture of the general findings. Sub-regional variability is high, particularly in areas with contrasted topography near coastland where also significant trend in precipitation variability and totals have been observed (e.g. Tu¨rkes° , 1996, 1998). The evaluation of regional data series (Fig. 63, right) indicate that the trend towards reduced winter precipitation trends in many regions are not statistically significant in view of the large variability (Xoplaki, 2002; Norrant and Dougue´droit, 2005). However, significant decreases are prevalent in western and central Mediterranean. For the Mediterranean Sea, precipitation variability and water budget have been investigated using gauge-satellite merged products and atmospheric re-analyses (Mariotti et al., 2002b). NCEP re-analyses show that during the last 50 years of the twentieth century Mediterranean averaged winter precipitation has decreased by about 20%, with the decrease mostly occurring during the period late 1970s to early 1990s. This implies a similar increase in the Mediterranean atmospheric water deficit with potentially important impacts on the Mediterranean Sea circulation as stressed in other chapters of this book. Xoplaki et al. (2004) in Fig. 66 show the monthly time evolution of the spatially averaged precipitation anomalies both for the instrumental data (300 stations equally distributed around the Mediterranean; upper panel) and the NCEP/NCAR re-analysis data (lower panel); their 4-year moving average low pass filtered time series are also shown to aid comparison at longer time scales. There is good agreement between the NCEP and station data (correlation 0.78). The plot highlights the underestimation of variability by the NCEP/ NCAR reanalysis data: the ratio of variance of observations vs. NCEP is 3.11. Decadal changes show good agreement between both datasets in depicting relatively wet and dry periods (see filtered data): relative maxima take place in the early 1950s, 1960s, late 1970s to early 1980s and late 1990s while relative minima occur in the late 1950s, early 1970s and early 1990s. The decadal changes are superimposed upon a long-term negative trend of 2.18 mm month 1 decade 1 (station data; 1.5 mm month 1 decade 1 for the reanalysis), significant at the 0.05 level. Naturally, negative trends of winter precipitation over western and central Mediterranean are probably linked with trends in cloud cover. In fact, recent studies for Portugal (Santos et al., 2002) and Italy (Maugeri et al., 2001), both using station-based data, have detected negative trends of winter cloud cover for the second part of the twentieth century. The results show that there is a highly significant negative trend in total cloud amount all over Italy. It is evident in all seasons and is particularly steep in winter where both in northern and southern Italy, the decrease exceeds 1 okta in 50 years (Maugeri et al., 2001).
204 Mediterranean Climate Variability
Figure 66: Upper panel: Wet season (October–March) precipitation anomalies (with respect to the 1950–1999 reference period) averaged over 292 sites in the Mediterranean (dashed line) and 4-year low pass filter (solid line); Lower panel: Wet season precipitation anomalies averaged over NCEP reanalysis data (dashed line) and 4-year low pass filter (solid line) (Adapted from Xoplaki et al., 2004).
3.5.3. Contributing Factors for Observed Temperature and Precipitation Trends After assessing the impact of several large-scale atmospheric circulation patterns in terms of Mediterranean climate, it is natural to expect that any major trends in the intensity of these circulation modes may impact directly on the observed climate. This is particularly true for the winter NAO index, that has revealed a strong positive trend throughout most of the 1980s and 1990s (Hurrell, 1995). However, over the last few winters a downward trend of the NAO has been
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
205
observed (Overland and Wang, 2005). The observed decreasing precipitation in northern Mediterranean and increasing Mediterranean fresh water deficit has been linked to trends of the NAO index in winter (Quadrelli et al., 2001; Mariotti et al., 2002b) influenced by corresponding trends in Atlantic storm-track paths and Mediterranean cyclogenesis (Trigo et al., 2000). On a monthly scale, probably the most significant precipitation trend can be found over the entire Iberian Peninsula for the month of March with a continuous monotonic decrease of up to 70% between the late 1950s and the 1990s (Trigo and DaCamara, 2000; Paredes et al., 2006). The spatial extent of those regions with significant trends (Fig. 67, top) is particularly coincident with the NAO impact region shown in Fig. 55. Moreover, the associated trend of cyclone densities for March shows a remarkable decrease near the Iberian Peninsula while northern Europe presents significant increase of cyclones and precipitation (Paredes et al., 2006). Recent studies have shown that throughout the latest two decades, the northern centre of the NAO dipole (the Icelandic low) has moved closer to Scandinavia (Jung and Hilmer, 2001). This shift has major implications for the Northern Hemisphere climate, in general, (Lu and Greatbatch, 2002) and for the precipitation field over Iberia, in particular (Rodo´ et al., 1997; Goodess and Jones, 2002). This observed trend in storm-track paths (Fig. 65, right) has been reinforced with observational studies showing a simultaneous coherent trend in blocking activity over the Atlantic sector (Barriopedro et al., 2005). It is not obvious if this variability is natural or induced itself by climate change. Recent works have shown that the significant winter precipitation decline that took place over the Mediterranean region during the last decades of the twentieth century results from the combined effect of trends of the NAO and the EA/ WRUS patterns (Krichak and Alpert, 2005a,b). These authors separated the Mediterranean basin in three target areas in order to identify the different trend effects of the NAO and EA/WRUS patterns. The area 7 E–10 E; 44 N–46 N was selected to represent the northwestern Mediterranean region. The target area for the northeastern Mediterranean region is located between 37 E–40 E; 35 N– 37 N, while that for the southeastern part of the basin is defined by 35 E–37 E; 31 N–34 N. Krichak and Alpert (2005a,b) based their analysis on a determination of the typical for winter months (DJF) circulation patterns over the northwestern, northeastern and southeastern Mediterranean area during the 15-year periods that are characterized by high and low NAO and EA/WRUS indices (1958–1972 and 1979–1993, respectively). The circulation patterns in the Figs. 68–70 illustrate the differences between the two 15-year periods. During the high phase period the typical for the northwestern part composite vortex is to be found positioned much further to the southeast (inland) than in the low phase case (Fig. 68). The difference explains the observed precipitation decrease during the last several decades of the past century by a decrease in the moisture content
Figure 67: Top: Decreasing Precipitation (DP) trends in March for the period 1941–1997. The different sizes of black dots depict the relative change in precipitation for the complete period after fitting March time series to a linear model. ‘‘Crosses’’ correspond to non-significant or positive trends, while the dots represent stations with declining precipitation at less than the 10% level (Mann–Kendall test). Bottom: Decadal trends (% relative to the mean over the study period) of the average number of cyclones detected in March for the period 1960–2000, computed on cell boxes with 10 longitude by 10 latitude and normalized for a standard latitude of 50 N. The solid line indicates the grid cells with significant trends at least at the 10% level (from Paredes et al., 2006).
A
B
Figure 68: Correlations between precipitation index over the northwestern Mediterranean target area and the u,v wind components at 850 hPa isobaric surface, during DJF months for (A) low six-year mean NAO-EA/WRUS; (B) high six-year mean NAO-EA/WRUS. The wind vectors represent the magnitudes of the wind-precipitation correlations obtained. The isolines and shaded areas in the figures represent the correlations and statistical significances (above 0.90) of the dependencies between relative vorticity and precipitation respectively (From Krichak and Alpert, 2005b).
208 Mediterranean Climate Variability
Figure 69: Same as in Fig. 68 but for the northeastern Mediterranean target area (From Krichak and Alpert, 2005b).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
209
A
B
Figure 70: Same as in Fig. 68 but for the southeastern Mediterranean target area. (From Krichak and Alpert, 2005b).
210 Mediterranean Climate Variability
of the air masses due to the stronger westerlies which characterize the high NAO months. Another physical mechanism appears to be linking the positive NAOEA/WRUS trend with the precipitation decline over the northeastern and southeastern Mediterranean regions (Figs. 69 and 70). The low-phase wet DJF months in the eastern Mediterranean (both northeastern and southeastern parts) are characterized by the circulation patterns with the northwesterly airflow (i.e. from the Atlantics towards the eastern Mediterranean) in the lower troposphere. One may expect that the moisture plays an important role in the intensification of the low troposphere northerly flows as well as stratospheric air intrusions and potential vorticity (PV) streamer systems over the southern Europe (Massacand et al., 1998). These processes significantly contribute to the cyclogenetic activity in the region. Additional analyses are required however for a better understanding of the physical mechanisms involved. On the contrary, the high phase years are characterized by a northeasterly airflow in the vicinity of the eastern Mediterranean. The increase in the role of the relatively dry continental air masses during the eastern Mediterranean wet months explains the precipitation decline over the eastern Mediterranean during the positive phase periods. Due to an evident relationship of the eastern Mediterranean synoptic processes with those associated with the development and consequent decay of the Asian–African monsoon (Webster et al., 1998) most of the precipitation of the southeastern Mediterranean region takes place during the cool season. The fact contributes to a noticeable focusing of the analyses on the winter-time processes. Figure 71 adapted from Xoplaki et al. (2004) presents standardized values of October–March precipitation anomalies over 110 Mediterranean stations for the 1900–1999 (RR 1900) period and the Gibraltar–Iceland NAO (Jones et al., 1997) with reversed sign. For the first half of the twentieth century, RR 1900 indicates the relatively dry early 1900s, early 1920s and 1940s and the wet periods 1910s and 1930s. RR 1900 suggests that the decreasing trend highlighted in Fig. 63 is not part of a centenial trend but a feature of the second half of the twentieth century and that the 1960s and late 1970s were actually the wettest intervals since the 1850s. This reasoning can be supported by the evolution of the large-scale circulation during the twentieth century shown in Fig. 71. The time series labelled as Cs1 (Fig. 71) shows the regressed time series (4-year moving average filter) between the NCAR SLP dataset (Trenberth and Paolino, 1980; 1900 to 1999) and the first canonical pair SLP pattern (presented in Fig. 62). Cs1 and RR 1900 show similar decadal changes since the beginning of the twentieth century (correlation 0.76). The correlation of NAO time series with Cs1 is 0.70, suggesting that the North Atlantic climate variability plays a crucial role in driving longterm trends in the Mediterranean. After (before) 1960, both Cs1 and -NAO show negative (positive) trends supporting the idea that the negative precipitation trends after the 1950s are dynamically induced and a feature of the second half
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
211
Figure 71: NAO: Gibraltar–Iceland NAO index (Jones et al., 1997) with reversed sign. RR 1900: spatial average over all Mediterranean sites with available data for the period 1900–1999. Cs1: regressed time series (4-years moving average filter) between the NCAR SLP dataset (Trenberth and Paolino, 1980) and the first canonical pair SLP pattern in Fig. 62 (Adapted from Xoplaki et al., 2004).
of the twentieth century. These results suggest that wet season Mediterranean precipitation increased since the second half of the nineteenth century and experienced a downward trend through the second half of the twentieth century (Xoplaki et al., 2004; see also Fig. 62). Further, the NAO index correlates at 0.72 with a large-scale Mediterranean Oscillation pattern during October– March (Du¨nkeloh and Jacobeit, 2003). The recently observed trend towards drier Mediterranean winter conditions is linked to particular circulation pattern changes including increased pressure south of 45 N–50 N since the 1970s, a weakening of the central Mediterranean trough since the late 1980s, and a long-term rising trend in the Mediterranean Oscillation pattern (higher pressure in western and central Mediterranean) being connected to the NAO (Du¨nkeloh and Jacobeit, 2003). Though the NAO mode plays an important role in driving temperature and precipitation trends in the Mediterranean, its influence varies through different time periods. Thus, there are other modes which are of relevance for explaining seasonal sub-Mediterranean climate variability (e.g. Kutiel and Benaroch, 2002; Kutiel et al., 2002; Du¨nkeloh and Jacobeit, 2003; Xoplaki et al., 2003a,b, 2004; Valero et al., 2004) or indirectly through the effect over sea level pressure patterns (Ribera et al., 2000). The nature of different rates of rainfall decrease in the east coast of Iberian Peninsula and parts of Italy (Brunetti et al., 2001, 2004) might be related to the observed increasing precipitation intensity, owing to a general enhancement of the hydrological cycle (caused by an increase in surface temperature), and the
212 Mediterranean Climate Variability
reduction of the number of wet days which are consistent with the variation of the atmospheric circulation (Brunetti et al., 2002, 2004). Minimum winter extreme temperatures across Peninsular Spain for the period 1955–1998 indicate that most of the extremes occurred under six synoptic patterns. A generalized decreasing trend in the annual frequency of extreme events is detected for most of the studied observatories. Prieto et al. (2004) showed that it is due to a non-linear shift in the annual mean minimum temperatures associated with a generalized warming in the area. An explanatory hypothesis of the differential diurnal warming observed over the western Mediterranean can be found in Fernandez-Garcia and Rasilla (2001). They showed an increase of the geopotential height over the region, particularly intense during the second half of the twentieth century. This was associated with increasing solar radiation, higher maximum temperatures and an intensified radiative loss at night, which have smoothed rising in daily minimum temperatures.
3.6. Other Important Forcing Factors 3.6.1. Tropical and Extratropical SST The direct influence of tropical climate variability on the Mediterranean region has long been debated and is discussed in Chapter 2 of this book. Here, the focus is on the link between mid-latitude circulation anomalies and tropical climate variability. In recent years, there has been a healthy and unclosed debate on the influence of the indirect effect, via extratropical modes, of El-Nin˜o-Southern Oscillation (ENSO) on climate patterns and precipitation in the Mediterranean region. A number of studies report that ENSO-related winter circulation anomalies in the Atlantic/European sector have a pattern similar to that of the NAO (Ribera et al., 2000; Pozo-Vazquez et al., 2001b; Mathieu et al., 2004). In particular, Pozo-Vazquez et al. (2001b) find that this is true but only for la Nin˜a events. There is still no general agreement on the extent of the ENSO influence in seasonal Mediterranean climate. Some studies suggest that the ENSO influences on the Mediterranean spring and autumn precipitation regime (Mariotti et al., 2002a), while others state that this influence is confined to the eastern Mediterranean in winter (Price et al., 1998). Rodo et al. (1997) instead find the ENSO signature over parts of southern Europe but only in areas where the NAO influence is weaker. However, others find no influence of ENSO on winter rainfall (Quadrelli et al., 2001). Some studies pointed to the large-scale changes induced by tropical processes and the role they might play in the modulation of either oceanic/atmospheric responses in certain mid-latitude areas. Of particular relevance appear to be the changes induced by the modulation of the local Hadley cell over the
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
213
Atlantic and the alteration of the ascending branch of this circulation, which affects the Mediterranean sea and in particular, the southern part. These dynamics have been traced by means of upper tropospheric humidity changes and varying heat fluxes (Fig. 72), as inferred from both cloud cover changes and absorbed solar radiation (Rodo´, 2001). Mariotti et al. (2002a, 2005) show an anomalously weaker (stronger) Azores anticyclone in connection with autumn (spring) rainfall anomalies in the Western Mediterranean. How these changes alter the net heat fluxes both in the tropical north Atlantic and the Mediterranean sea (particularly the western basin) and which might be their ultimate effects on Mediterranean climate is the subject of active research and current debate. A cooling of the western Mediterranean Sea appears to take place linked to a sequence of atmosphere–ocean couplings that initiate in the warm tropical Pacific during an El Nin˜o event (Klein et al., 1999; Rodo´, 2001). Several authors have suggested the role played by ocean dynamics (Sutton and Allen, 1997; Rodwell et al., 1999) as potential sources of variability affecting the North Atlantic climate with impacts on the Mediterranean regions. Hoerling et al. (2001) found that observed long-term changes in the NAO during winter since 1950 are recoverable from tropical SST forcing alone. Latif (2001) show some skill in predicting long-term changes in the NAO based on tropical SSTs. Recent numerical experiments by Hoerling et al. (2004) and Hurrell et al. (2004) indicate that tropical SST variations, particularly in tropical Indian and western Pacific Ocean, have significantly controlled recent North Atlantic circulation anomalies. While the underlying mechanisms behind these tropical–extratropical connection remain to be explained, these studies suggest a potential for predicting changes such as the dryness observed in the Mediterranean region in connection to the trend in the NAO based on tropical SST anomalies. Effects of other tropical systems like, Asian (Indian) Monsoon, African Monsoon, Hurricanes and Saharan dust on the Mediterranean climate were also reported in many studies and are summarized in Chapter 2 of this book.
3.6.2. Solar Variability Solar variability and troposphere–stratosphere interaction (Perlwitz and Graf, 1995, 2001; Shindell et al., 1999) have been recognized to impact the North Atlantic climate. Solar influence on the Earth’s climate is a long-discussed topic and it has produced controversial scientific opinions. In general terms the solar cycle relationship is viewed as just a statistical fluctuation, however the solar– climate relationship could be strongly nonlinear. An example of this nonlinearity is the problem of the spatial structure of the NAO according to the solar cycle.
214 Mediterranean Climate Variability
1
6
< -0.6 -0.4 -0.2 0.2 0.4 0.6 >
Figure 72: Correlation fields between an ENSO index and Upper-tropospheric humidity, at lags 1 and 6 months. Contoured areas indicate a significance higher than p < 0.001. Continuous lines indicate positive values and dotted line refers to negative ones. (Adapted from Rodo´, 2001).
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
215
Kodera (2002, 2003) demonstrated that the spatial structure of the NAO during the winter is very different during low solar activity (NAO confined in the Atlantic sector) than during high solar activity (hemispherical structure extending into the stratosphere). This difference has important consequences as the modulation of the winter and summer circulation linkage (Ogi et al., 2003) or the modulation of the relationship between NAO and the northern hemisphere surface air temperature (Gimeno et al., 2003). Very recently, Kodera and Kuroda (2005) have proposed a possible mechanism to explain the solar modulation of the spatial structure of the NAO. They suggest that the solar activity influence originates in the stratopause region from a change in the seasonal march of the jet. The leading mode of the interannual variation of the zonal–mean zonal wind in the stratopause region has a meridional dipole-type anomaly structure in high solar activity winters. This structure extends into the troposphere by changes in the meridional propagation of planetary waves giving a hemispherical structure to the NAO. During low solar activity, the downward extension of the zonal–mean zonal wind anomalies is weak, so regional scale variations are dominant in the troposphere and NAO is confined in the Atlantic sector. It has been shown that when the long-term solar activity is high, then the smoothed NAO index is low and vice-versa (Kirov and Georgieva, 2002). Is the strength of the NAO–solar connection sufficiently strong to impact directly the Mediterranean climate? We believe that it is. A recent study on the frequency and magnitude of flood episodes in Iberian rivers over the last millennium shows that periods of high frequency in floods are well associated with periods of high solar activity (Vaquero, 2004).
3.7. Future Outlook In the twentieth century the Mediterranean was characterized by positive trends of temperature in every season but particularly in winter and summer. However, if one restricts the analysis to the last three or four decades, the positive trend is prominent in the summer (particularly in the west), while areas of negative trend can be observed in spring. On the other hand, the twentieth century precipitation regime was characterized by negative (positive) winter precipitation trends over the eastern (western) Mediterranean regions, while autumn and spring reveal mostly negative trends in Iberia and central Mediterranean regions. Over the last 50 years, there has been a significant decrease of precipitation in winter and spring, with a conspicuous decline of March precipitation throughout the western Mediterranean sector. The physical processes responsible for these trends and changes seem to be partially of a hemispheric nature (such as external forcings and changes in the
216 Mediterranean Climate Variability
large-scale atmospheric circulation) as well as local/regional (such as changes in earth surface and land use). In fact, temperature and precipitation trends over the Mediterranean, during the twentieth century, are shown to be directly linked to changes in phase of some of the major large-scale circulation modes previously mentioned, particularly for NAO and EA/WRUS patterns during the winter season. It should be stressed that most studies on variability and trends of precipitation over the Mediterranean region and associated atmospheric circulation patterns focus on the winter part of the year. Some of these studies use the standard winter 3 months (DJF), others also incorporate the months of March (e.g. Trigo et al., 2002b, 2004a), some even use the wet season concept between October and March (e.g. Xoplaki et al., 2004). In any case, the highly seasonal behaviour of the precipitation regime is reflected in a bias towards winter on the number of scientific works published. We acknowledge that such bias is also present in this chapter. Studies on trends of temperature cover more evenly the various seasons of the year, with slightly more focus on the two extreme seasons, i.e. summer and winter. As we have shown in this chapter, considerable research has been carried out produced in recent years linking the most relevant large-scale atmospheric circulation modes with Mediterranean climate variables. In particular, it is now clear that twentieth century trends of precipitation, temperature, cloud cover can be attributed, at least partially, to corresponding trends of well-established circulation patterns. Nevertheless, there are still grey areas, and different issues require further analysis in the near future. Here we provide a list of issues that in our view should be dealt explicitly during the next decade: (1) There are large gaps between datasets for the northern and southern margins of the Mediterranean basin. Therefore, it is necessary to assemble more reliable and consistent climate datasets (both at the daily and monthly scale) over the entire Mediterranean basin. Among other initiatives, it would be particularly helpful to contribute to the production of a high resolution regional reanalyses dataset. (2) Despite the large amount of work already done on climate trends, we acknowledge that there is still scope for further work. Therefore, it is necessary to confirm the role played by important large-scale modes to explain regional temperature and precipitation variability and trends for different periods and seasons within the instrumental period using sophisticated statistical methods. It is necessary to characterize the climatic impact imposed by these atmospheric circulation modes on a month-by-month basis and to assess precisely where these impacts are significant. Furthermore,
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
217
especial attention should be devoted to characterize changes of the annual cycle in Mediterranean temperature and precipitation through the instrumental period. (3) Climate change scenarios obtained with different General Circulation Models (GCMs) reveal that there is a general agreement towards a significant decrease of precipitation for this region among most GCMs (Folland et al., 2001). In fact, this is the region of the planet that shows a more consistent signal towards a dry future as shown by the analysis of climate model simulations regarding the reproducibility of the main characteristics of the extra tropical modes, including its impact on the precipitation field and cyclonic activity (Osborn et al., 1999; Ulbrich and Christoph, 1999; Trigo and Palutikof, 2001). However, present climate models are still unable to replicate the observed amplitude of the interannual variability and of the multidecadal trends of some modes, e.g. the NAO (Osborn, 2004). Therefore, this work should be continued and intensified with the aim of determining to what extent climate models yield a realistic picture of the variability in the present climate and quantifying, if possible, the amount of expected regional climate change that can be ascribed to future trends in extra tropical modes, since these modes will probably be responsible for regional differences in the future climate. (4) Trends in the large-scale driving patterns are especially relevant since they may enhance or dampen the warming caused by increasing greenhouse gas concentrations. However, to our knowledge, work has just begun in this direction. Relevant questions in this context are the possible changes of these extra tropical modes under global climate change. The results to date seem to indicate that the so-called Annular Modes – the Arctic Oscillation and the Antarctic Oscillation – to which the NAO is linked will tend to become more intense in the future (Guillet et al., 2000), although the signal-to-noise ratio may be not very large (Zorita and Gonza´lez-Rouco, 2000). In a more regional basis, some work has been reported for the Western Mediterranean (Gonza´lez-Rouco et al., 2000) and the Balkans (Busuioc et al., 1999). These issues are further stressed in Chapter 8 of this book. (5) Many studies have underlined the role of several teleconnection patterns (NAO, EA/WRUS, EA, EA-Jet, SCAND) over different parts of the Mediterranean basin. However, in general it is not clear whether there are significant differences in the physical mechanisms for the generation and dissipation of these planetary regimes, or whether the only difference lies simply in the geographical location of their centres of action. Furthermore, the underlying physical mechanisms (moisture and enthalpy advection,
218 Mediterranean Climate Variability
cyclogenesis and storm tracks, vertical stability, radiation and cloud cover, oceanic processes etc.) that give rise to these statistical connections are not always completely understood. This question is three-fold relevant: in the context of changes in the intensity of these circulation patterns under global climate change. Only with a sufficient understanding of the physical mechanisms will it be possible to estimate the effect of changes of extra tropical modes on the Mediterranean climate. For understanding the variability at decadal time scales (where persistent deviations of atmospheric indices have been observed), and estimate their influence on low-frequency changes in temperature and precipitation, in the frame of other possibly competing effects, such as land use changes. Seasonal climate prediction would greatly benefit from research in this direction. A substantial fraction of the climate variability of the Mediterranean basin may be explained in terms of a few planetary flow regimes. Their predictability, though, is not well quantified, and their origin and possible relationship with oceanic process is very uncertain. Possibly, in order to make accurate predictions of variables like precipitation, some kind of downscaling method would be necessary to fill the gap between the large and the regional scales.
Acknowledgements The authors are indebted to Dr. Clare Goodess for her comments and suggestions that helped to improve the clarity of this chapter. Ricardo Trigo and Isabel Trigo (CGUL) were supported by the Portuguese Science Foundation (FCT) through project VAST (Variability of Atlantic Storms and their impact on land climate) Contract POCTI/CTA/46573/2002, cofinanced by the European Union under program FEDER. Ju¨rg Luterbacher and Elena Xoplaki are supported by the Swiss National Science Foundation (NCCR). Elena Xoplaki was financially supported through the European Environment and Sustainable Development programme, projects SOAP (EVK2-CT-2002-00160) and EMULATE (EVK2-CT-2002-00161). Maurizio Maugeri (Milan University) was supported by the project CLIMAGRI (Italian Ministry for agriculture and forests). Michele Brunetti and Teresa Nanni (ISAC-CNR) were supported by the ALP-IMP project (EU-FP5) and U.S.– ITALY bilateral Agreement on Cooperation in Climate Change Research and Technology (Italian Ministry for the environment). Jucundus Jacobeit was supported by the EU project EMULATE (European and North Atlantic daily
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
219
to multidecadal climate variability), EVK2-CT-2002-00161. Jesu´s Fernandez was funded by the Basque Regional Government through grant BFI04.52.
References Barnston, A. G., & Livezey, R. E. (1987). Classification, seasonality and persistence of low-frequency atmospheric circulation patterns. Mon. Wea. Rev., 115(1–2), 1083–1127. Barriopedro, D., Garcı´ a-Herrera, R., Lupo, A. R., & Herna´ndez, E. (2005). A climatology of Northern Hemisphere blocking, J. Climate, (accepted). Barry, R. G., & Chorley, R. J. (2003). Atmosphere, weather and climate. Routledge, London). Ben-Gai, T., Bitan, A., Manes, A., Alpert, P., & Kushnir, Y. (2001). Temperature and surface pressure anomalies in Israel and the North Atlantic Oscillation. Theor. Appl. Climatol., 69(3–4), 171–177. Bro¨nnimann, S., & Luterbacher, J. (2004). Reconstructing Northern Hemisphere upperlevel fields during World War II. Clim. Dyn., 22, 499–510. Brunet, M., Aguilar, E., Saladı´ e, O., Sigro´, J., & Lo´pez, D. (2001). The variations and trends of the surface air temperature in the Northeastern Spain from middle nineteenth century onwards. In: M. Brunet, and D. Lo´pez, (Eds), Detecting and Modelling Regional Climate change, (Berlin, Springer). Brunetti, M., Colacino, M., Maugeri, M., & Nanni, T. (2001). Trends in the daily intensity of precipitation in Italy from 1951 to 1996. Int. J. Climatol., 21(D18), 299–316. Brunetti, M., Maugeri, M., & Nanni, T. (2002). Atmospheric circulation and precipitation in Italy for the last 50 years. Int. J. Climatol., 22, 1455–1471. Brunetti, M., Maugeri, M., Monti, F., & Nanni, T. (2004). Changes in daily precipitation frequency and distribution in Italy over the last 120 years. J. Geophys. Res. Atmosph., 109, D05, doi:10.1029/2003JD004296. Brunetti, M., Maugeri, M., Monti, F., & Nanni, T. (2006). Temperature and precipitation variability in Italy in the last two centuries from homogenized instrumental time series. Int. J. Climatol., in press. Busuioc, A., von Storch, H., & Schnur, R. (1999). Verification of GCM generated regional precipitation and of statistical downscaling estimates. J. Climate, 12, 258–272. Castro-Dı´ ez, Y., Pozo-Va´zquez, D., Rodrigo, F. S., & Esteban-Parra, M. J. (2002). NAO and winter temperature variability in southern Europe,. Geophys. Res. Lett., 29, 8, doi: 10.1029/2001GL014042. Conte, M., Giuffrida, S., & Tedesco, S. (1989). The Mediterranean oscillation: impact on precipitation and hydrology in Italy. Proceedings of the conference on climate and water, Publications of the Academy of Finland, Helsinki, 1, 121–137. Corte-Real, J., Zhang, X., & Wang, X. (1995). Large-scale circulation regimes and surface climatic anomalies over the Mediterranean. Int. J. Climatol., 15, 1135–1150. Cullen, H. M., & de Menocal, P. B. (2000). North Atlantic influence on Tigris–Euphrates streamflow. Int. J. Climatol., 20, 853–863. Du¨nkeloh, A., & Jacobeit, J. (2003). Circulation dynamics of Mediterranean Precipitation Variability 1948–1998. Int. J. Climatol., 23, 1843–1866. Ferna´ndez, F., & Rasilla, D. (2001). Secular variations of the synoptic circulation over the Iberian Peninsula. In: M. Brunet, and D. Lo´pez, (Eds), Detecting and Modelling Regional Climate Change, (Berlin, Springer).
220 Mediterranean Climate Variability
Ferna´ndez, J., Sa´enz, J., & Zorita, E. (2003). Analysis of wintertime atmospheric moisture transport and its variability over southern Europe in the NCEP-Reanalyses. Clim. Res., 23, 195–215. Folland, C. K., et al. (2001). Observed climate variability and change, in Chapter 2 of climate change 2001; Houghton, J. T. et al. (Eds), the scientific basis, contribution of working group I to the third assessment report of the Intergovernmental Panel on Climate Change, Cambridge University Press, Cambridge, UK and New York, NY, USA. Galan, E., Can˜ada, R., Ferna´ndez, F., & Cervera, B. (2001). Annual temperature evolution in the southern plateau of Spain from the construction of regional climatic time series. In: M. Brunet, and S. Lo´pez (Eds), Detecting and Modelling Regional Climate Change, (Springer-Verlag, Berlin, pp. 119–131). Gimeno, L., de la Torre, L., Nieto, R., Garcı´ a, R., Herna´ndez, E., & Ribera, P. (2003). Changes in the relationship NAO-Northern Hemisphere temperature due to solar activity. Earth Planet. Sci. Lett., 206, 15–20. Giorgi, F. (2002). Variability and trends of sub-continental scale surface climate in the twentieth century. Part I: observations. Clim. Dyn., 18, 675–791. Gonza´lez-Rouco, J. F., Heyen, H., Zorita, E., & Valero, F. (2000). Agreement between observed rainfall trends and climate change simulations in the southwest of Europe. J. Climate, 13, 976–985. Gonza´lez-Rouco, J. F., Jime´nez, J. L., Quesada, V., & Valero, F. (2001). Quality control and homogenization of monthly precipitation data in the southwest of Europe. J. Climate, 14, 964–978. Goodess, C. M., & Jones, P. D. (2002). Links between circulation and changes in the characteristics of Iberian rainfall. Int. J. Climatol., 22, 1593–1615. Guillet, N., Hegerl, G. C., Allen, M. R., & Scott, P. A. (2000). Implications of changes in the Northern Hemisphere circulation for the detection of anthropogenic climate change. Geophys. Res. Lett., 27, 993–996. Hansen, J. E., Ruedy, R., Sato, M. K., Imhoff, M., Lawrence, W., Easterling, D., Peterson, T., & Karl, T. (2001). A closer look at United States and global surface temperature change. J. Geophys. Res., 106, 23947–23963. Hasanean, H. M. (2004). Wintertime surface temperature in Egypt in relation to the associated atmospheric circulation. Int. J. Climatol., 24, 985–999. Hess, P., & Brezowski, H. (1977). Katalog der Grosswetterlagen Europas (1881–1976). Berichte des Deutschen Wetterdienstes 113(15), Selbstverlag des Deutschen Wetterdienstes, Offenbach am Main, Germany. H. M. S. O. (1962). Weather in the Mediterranean I: general meteorology. 2nd ed., London, 362 p. Hoerling, M. P., Hurrell, J. W., & Xu, T. (2001). Tropical origins for recent North Atlantic climate change. Science, 292, 90–92. Hoerling, M. P., Hurrell, J. W., Xu, T., Bates, G. T., & Phillips, A. S. (2004). Twentieth Century North Atlantic Climate change. Part II: understanding the effect of Indian Ocean warming. Clim. Dyn., 23, 391–405. Hurrell, J. W. (1995). Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitation. Science, 269, 676–679. Hurrell, J. W., & van Loon, H. (1997). Decadal variations in climate associated with the North Atlantic Oscillation. Clim. Change, 36, 301–326.
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
221
Hurrell, J. W., Hoerling, M. P., Phillips, A. S., & Xu, T. (2004). Twentieth Century North Atlantic Climate change. Part I: assessing determinism. Climate dynamics. Clim. Dyn., 23, 371–389. Jacobeit, J. (2000). Rezente Klimaentwicklung im Mittelmeerraum. Petermanns Geographische Mitteilungen, 144, 22–33. Jacobeit, J., Du¨nkeloh, A., & Hertig, E. (2004). Die Niederschlagsentwicklung im mediterranen Raum und ihre Ursachen. In: J. L. Lozan, H. Grassl, P. Hupfer, L. Menzel, & C. D. Scho¨nwiese, (Eds), Warnsignal Klima: Genug Wasser fu¨r alle?, (Hamburg, pp. 192–196). Jones, P. D., Jonsson, T., & Wheeler, D. (1997). Extension to the North Atlantic Oscillation using instrumental pressure observations from Gibraltar and south-west Iceland. Int. J. Climatol., 17, 1433–1450. Jung, T., & Hilmer, M. (2001). On the link between the North Atlantic Oscillation and Arctic sea ice export through Fram Strait. J. Climate, 14, 3932–3943. Kimoto, M., & Ghil, M. (1993). Multiple flow regimes in the Northern Hemisphere winter. Part II: sectorial regimes and preferred transitions. J. Atmos. Sci., 50, 2645–2673. Kirov, B., & Georgieva, K. (2002). Long term variations and interrelations of ENSO, NAO & Solar Activity. Phys. Chem. Earth., 27, 441–448. Klein, S. A., Soden, B. J., & N.-Lau, C. (1999). Remote sea surface temperature variations during ENSO: evidence for a tropical atmospheric bridge. J. Clim., 12, 917–932. Kodera, K. (2002). Solar cycle modulation of the North Atlantic Oscillation: Implication in the spatial structure of the NAO. Geophys. Res. Lett., 29, 1218, doi: 10.1029/ 2001GL014557. Kodera, K. (2003). Solar influence on the spatial structure of the NAO during the winter 1900–1999. Geophys. Res. Lett., 30, 1175, doi: 10.1029/2002GL016584. Kodera, K., & Kuroda, Y. (2005). A possible mechanism of solar modulation of the spatial structure of the North Atlantic Oscillation. J. Geophys. Res., 110, D02111, doi: 10.1029/2004JD005258. Ko¨ppen, W. (1936). In: Das geographische System der Klimate: Handbuch der Klimatologie. W. Ko¨ppen, and R. Geiger, (Eds), Vol. 3, (Gebru¨der Borntra¨ger, Berlin, p. 46). Krichak, S. O., Kishcha, P., & Alpert, P. (2002). Decadal trends of main Eurasian oscillations and the Eastern Mediterranean precipitation. Theor. Appl. Climatol., 72, 209–220. Krichak, S. O., & Alpert, P. (2005a). Decadal trends in the east Atlantic/West Russia pattern and the Mediterranean precipitation. Int. J. Climatol., 25, 183–192. Krichak, S. O., & Alpert, P. (2005b). Signatures of the NAO in the atmospheric circulation during wet winter months over the Mediterranean region. Theor. Appl. Climatol., (In press). Kutiel, H., & Benaroch, Y. (2002). North Sea-Caspian Pattern (NCP) – an upper level atmospheric teleconnection affecting the eastern mediterranean: Identification and definition. Theor. Appl. Climatol., 71, 17–28. Kutiel, H., Maheras, P., Tu¨rkes° , M., & Paz, S. (2002). North Sea – Caspian Pattern (NCP) – an upper level atmospheric teleconnection affecting the eastern Mediterranean – implications on the regional climate. Theor. Appl. Climatol., 72, 173–192.
222 Mediterranean Climate Variability
Lamb, P. (1972). British Isles weather types and a register of daily sequence of circulation patterns, 1861–1971, Geophysical Memoir 116, HMSO, London, 85 pp. Lamb, P., & Peppler, R. (1987). The North Atlantic Oscillation: concept and an application. Bull. Amer. Meteor. Soc., 68, 1218–1225. Latif, M. (2001). Tropical Pacific/Atlantic Ocean interactions at multi-decadal time scales. Geophys. Res. Lett., 28, 535–542. Lu, J., & Greatbatch, R. J. (2002). The changing relationship between the NAO and the Northern Hemisphere Climate Variability. Geophys. Res. Lett., 29, 10, doi: 1029/2001GLO14052. Luterbacher, J., Xoplaki, E., Dietrich, D., Rickli, R., Jacobeit, J., Beck, C., Gyalistras, D., Schmutz, C., & Wanner, H. (2002). Reconstruction of sea-level pressure fields over the Eastern North Atlantic and Europe back to 1500. Clim. Dyn., 18, 545–561. Mariotti, A., Zeng, N., & Lau, K. M. (2002a). Euro-Mediterranean rainfall and ENSO – a seasonally varying relationship. Geophys. Res. Lett., 29, 12, doi:10.1029/ 2001GL014248. Mariotti, A., Struglia, M. V., Zeng, N., & K.-Lau, M. (2002b). The hydrological cycle in the Mediterranean region and implications for the water budget of the Mediterranean Sea. J. Climate, 15, 1674–1690. Mariotti, A., Ballabrera-Poy, J., & Zeng, N. (2005). Tropical influence on EuroAsian autumn rainfall. Clim. Dyn., 24(5), 10.1007/s00382-004-0498-6: 511–521. Massacand, A. C., Wernly, H., & Davies, H. C. (1998). Heavy precipitation on the Alpine south-side: an upper-level precursor. Geophys. Res. Lett., 25, 1435–1438. Mathieu, P., Sutton, R. T., Dong, B., & Collins, M. (2004). Predictability of winter climate over the North Atlantic European region during ENSO events. J. Climate, 17, 1953–1974. Maugeri, M., Bagnati, Z., Brunetti, M., & Nanni, T. (2001). Trends in Italian total cloud amount, 1951–1996. Geophys. Res. Lett., 28, 4551–4554. New, M. G., Hulme, M., & Jones, P. D. (2000). Representing twentieth-century space-time climate variability. Part II: development of 1901–96 monthly grids of terrestrial surface climate. J. Climate, 13, 2217–2238. New, M., Todd, M., Hulme, M., & Jones, P. D. (2001). Precipitation measurements and trends in the twentieth century. Int. J. Climatol., 21, 1899–1922. NOAA-CPC (2005a). Northern Hemisphere Teleconnection Patterns. http:// www.cpc.ncep.noaa.gov/data/teledoc/telecontents.html.050405. NOAA-CPC (2005b). Calendar months when specific teleconnection patterns are important. http://www.cpc.ncep.noaa.gov/data/teledoc/teletab.gif. Norrant, C., & Dougue´droit (2005). Monthly and daily precipitation trends in the Mediterranean (1950–2000). Theor. Appl. Climatol., (in press). Ogi, M., Yamazaki, K., & Tachibana, Y. (2003). Solar cycle modulation of the seasonal linkage of the North Atlantic Oscillation (NAO). Geophys. Res. Lett., 30, 2170, doi:10.1029/2003GL018545. Osborn, T. J., Briffa, K. R., Tett, S. F. B., Jones, P. D., & Trigo, R. M. (1999). Evaluation of the North Atlantic Oscillation as simulated by a climate model. Clim. Dyn., 15, 685–702. Osborn, T. J. (2004). Simulating the winter North Atlantic Oscillation: the roles of internal variability and greenhouse gas forcing. Clim. Dyn., 22, 605–623.
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
223
Overland, J. E., & Wang, M. (2005). The Artic climate paradox: the recent decrease of the Artic Oscillation. Geophys. Res. Lett., 32, L06701, doi:10.1029/ 2004GL021752. Paredes, D., Trigo, R. M., Garcia-Herrera, R., & Trigo, I. F. (2005). Understanding precipitation changes in Iberia in early Spring: weather typing and storm-tracking approaches, Journal of Hydrometeorology, 7, 101–113. Perlwitz, J., & Graf, H. F. (1995). The statistical conection between tropospheric and stratospheric circulation of the Northern Hemisphere in winter. J. Clim., 8, 2281–2295. Perlwitz, J., & Graf, H. F. (2001). Troposphere-stratosphere dynamic coupling under strong and weak polar vortex condition. Geophys. Res. Lett., 28, 271–274. Pozo-Va´zquez, D., Esteban-Parra, M. J., Rodrigo, F. S., & Castro-Diez, Y. (2001a). The association between ENSO and winter atmospheric circulation and temperature in the North Atlantic region. J. Climate, 14, 3408–3420. Pozo-Va´zquez, D., Esteban-Parra, M. J., Rodrigo, F., & Castro-Dı´ ez, Y. (2001b). A study of NAO variability and its possible non-linear influences on European surface temperature. Clim. Dyn., 17, 701–715. Price, C., Stone, L., Rajagopalan, B., & Alpert, P. (1998). A possible link between El Nino and precipitation in Israel. Geophys. Res. Lett., 25, 3963–3966. Prieto, L., Garcı´ a-Herrera, R., Dı´ az, J., Herna´ndez, E., & del Teso, M. T. (2004). Minimum Extreme Temperatures over Peninsular Spain. Global and Planetary Change, 44, 59–71, doi:10.1016/j.gloplacha.2004.06.005. Qian, B., Corte-Real, J., & Xu, H. (2000). Is the North Atlantic Oscillation the most important atmospheric pattern for precipitation in Europe? J. Geophys. Res., 105, 11901–11910. Quadrelli, R., Pavan, V., & Molteni, F. (2001). Wintertime variability of Mediterranean precipitation and its links with large-scale circulation anomalies. Clim. Dyn., 17, 5–6, 457–466. Rex, D. F. (1950a). Blocking action in the middle troposphere and its effect upon regional climate. Part I.: An aerological study of blocking action. Tellus, 2, 196–211. Rex, D. F. (1950b). Blocking action in the middle troposphere and its effect upon regional climate. Part II: The climatology of blocking action. Tellus, 2, 275–301. Rex, D. F (1951). The effect of Atlantic blocking action upon European climate. Tellus, 3, 1–16. Ribera, P., Garcia, R., Diaz, H. F., Gimeno, L., & Hernandez, E. (2000). Trends and interannual oscillations in the main sea-level surface pressure. patterns over the Mediterranean, 1955–1990. Geophys. Res. Lett., 27, 1143–1146. Rodo´, X., Baert, E., & Comin, F. A. (1997). Variations in seasonal rainfall in Southern Europe during the present century: relationships with the North Atlantic Oscillation and the El Nin˜o-Southern oscillation. Clim. Dyn., 13, 275–284. Rodo´, X. (2001). Reversal of three global atmospheric fields linking changes in SST anomalies in the Pacific, Atlantic and Indian oceans at tropical latitudes and midlatitudes. Clim. Dyn., 18, 203–217. Rodrı´ guez-Fonseca, B., & de Castro, M. (2002). On the connection between winter anomalous precipitation in the Iberian Peninsula and North West Africa and the summer subtropical atlantic sea surface temperature. Geophys. Res. Lett., 29, 10.1029/ 2001GL014421. Rodwell, M. J., Rowell, D. P., & Folland, C. K. (1999). Oceanic forcing of the wintertime North Atlantic Oscillation and European Climate. Nature, 398, 320–323.
224 Mediterranean Climate Variability
Rogers, J. C. (1984). The association between the North Atlantic Oscillation and the Southern Oscillation in the Northern Hemisphere. Mon. Wea. Rev., 112, 1999–2015. Ro¨tter, R., & van de Geijn, S. C. (1999). Climate change effects on plant growth, crop yield and livestock. Clim. Change, 43, 651–681. Sa´enz, J., Zubillaga, J., & Rodriguez-Puebla, C. (2001a). Interannual variability of winter precipitation in northern Iberian Peninsula. Int. J. Climatol., 21, 1503–1513. Sa´enz, J., Rodrı´ guez-Puebla, C., Ferna´ndez, J., & Zubillaga, J. (2001b). Interpretation of interanual winter temperature variations over southwestern Europe. J. Geophys. Res., 106, 20641–20652. Santos, F. D., Forbes, K., & Moita, R. (2002). Climate change in Portugal. Scenarios, impacts and adaptation measures – SIAM Project, Gradiva, Lisbon. Schmutz, C., Luterbacher, J., Gyalistras, D., Xoplaki, E., & Wanner, H. (2000). Can we trust proxy-based NAO index reconstructions? Geophys. Res. Lett., 27, 1135–1138. Scho¨nwiese, C. D. (1993). Klimatrend-Atlas Europa 1891–1990 (Atlas of climate trends for Europe 1891–1990). Verlag, Zentrum fu¨r Umweltforschung, Frankfurt. 218 pp. Shindell, D. T., Miller, R. L., Schmidt, G., & Pandolfo, L. (1999). Simulation of the Arctic oscillation trend by greenhouse forcing of a stratospheric model. Nature, 399, 453–455. Sigro´, J., (2004). Variabilidad espacio-temporal de la temperatura del aire en Catalun˜a, Ph.D, University of Rovira i Virgili, Tarragona. Struglia, M. V., Mariotti, A., & Filograsso, A. (2004). River discharge into the Mediterranean Sea: climatology and aspects of the observed variability. J. Climate, 17, 4740–4751. Sutton, R. T., & Allen, M. R. (1997). Decadal predictability of North Atlantic sea surface temperature and climate. Nature, 388, 563–567. Thompson, D. W. J., & Wallace, J. M. (1998). The Arctic Oscillation signature in the wintertime geopotential height and temperature fields. Geophys. Res. Lett., 25, 1297–1300. Tibaldi, S., D’Andrea, F., Tosi, E., & Roeckner, E. (1997). Climatology of Northern Hemisphere blocking in the ECHAM model. Clim. Dyn., 13, 649–666. Touchan, R., Xoplaki, E., Funkhouser, G., Luterbacher, J., Hughes, M. K., Erkan, N., Akkemik, U¨., & Stephan, J. (2005). Reconstructions of Spring/Summer precipitation for the Eastern Mediterranean from tree-ring widths and its connection to large-scale atmospheric circulation. Clim. Dyn., DOI:10.1007/s00382s-005–0016-5. Trenberth, K., & Paolino, D. A. (1980). The Northern Hemisphere sea level pressure data set: trends, errors and discontinuities. Mon. Weather Rev., 108, 855–872. Trigo, I. F., Davies, T. D., & Bigg, G. R. (1999). Objective climatology of cyclones in the Mediterranean region. J. Climate, 12, 1685–1696. Trigo, I. F., Davies, T. D., & Bigg, G. R. (2000). Decline in Mediterranean rainfall caused by weakening of Mediterranean cyclones. Geophys. Res. Lett., 27, 2913–2916. Trigo, I. F., Bigg, G. R., & Davies, T. D. (2002a). Climatology of cyclogenesis mechanisms in the Mediterranean. Mon. Wea. Rev., 130, 549–569. Trigo, R. M., & DaCamara, C. C. (2000). Circulation weather types and their impact on the precipitation regime in Portugal. Int. J. of Climatol., 20, 1559–1581. Trigo, R. M., & Palutikof, J. P. (2001). Precipitation scenarios over Iberia: a comparison between direct GCM output and different downscaling techniques. J. Climate, 14, 4422–4446.
Relations between Variability in the Mediterranean Region and Mid-Latitude Variability
225
Trigo, R. M., Osborn, T. J., & Corte-Real, J. (2002b). The North Atlantic Oscillation influence on Europe: climate impacts and associated physical mechanisms. Clim. Res., 20, 9–17. Trigo, R. M., Pozo-Va´zquez, D., Osborn, T. J., Castro-Dı´ ez, Y., Ga´miz-Fortis, S., & Esteban-Parra, M. J. (2004a). North Atlantic Oscillation influence on precipitation, river flow and water resources in the Iberian Peninsula. Int. J. Climatol., 24, 925–944. Trigo, R. M., Trigo, I. F., DaCamara, C. C., & Osborn, T. J. (2004b). Climate impact of the European winter blocking episodes from the NCEP/NCAR reanalyses. Clim. Dyn., 23, 17–28. Tu¨rkes° , M. (1996). Spatial and temporal analysis of annual rainfall variations in Turkey. Int. J. Climatol., 16, 1057–1076. Tu¨rkes° , M. (1998). Influence of geopotential heights, cyclone frequency and southern oscillation on rainfall variations in Turkey. Int. J. Climatol., 18, 649–680. Tu¨rkes° , M., & Erlat, E. (2003). Precipitation changes and variability in Turkey linked to the North Atlantic Oscillation during the period 1930–2000. Int. J. Climatol., 23, 1771–1796. Tu¨rkes° , M., & Erlat, E. (2005). Climatological responses of winter precipitation in Turkey to variability of the North Atlantic Oscillation during the period 1930–2001. Theor. Appl. Climatol., Online paper: ISSN: 0177–798, 1434–4483. _ (2002). Re-evaluation of trends and changes Tu¨rkes° , M., Su¨mer, U. M., & Demir, I. in mean, maximum and minimum temperatures of Turkey for the period 1929–1999. Int. J. Climatol., 22, 947–977. Tu¨rkes° , M., & Su¨mer, U. M. (2004). Spatial and temporal patterns of trends and variability in diurnal temperature ranges of Turkey. Theor. Appl. Climatol., 77, 195–227. Ulbrich, U., & Christoph, M. (1999). A shift in the NAO and increasing storm track activity over Europe due to anthropogenic greenhouse gas. Clim. Dyn., 15, 551–559. Ulbrich, U., Christoph, M., Pinto, J. G., & Corte-Real, J. (1999). Dependence of winter precipitation over Portugal on NAO and baroclinic wave activity. Int. J. Climatol., 19, 379–390. van Loon, H., & Rogers, J. C. (1978). The seesaw in winter temperatures between Greenland and Northern Europe, Part 1: general description. Mon. Wea. Rev., 106, 296–310. Valero, F., Luna, M. Y., Martin, M. L., Morata, A., & Gonza´lez-Rouco, J. F. (2004). Coupled modes of large scale climate variables and regional precipitation in the Western Mediterranean in autumn. Clim. Dyn., 22, 307–323. Vaquero, J. (2004). Solar signal in the number of floods recorded for the Tagus river basin over the last millennium. Climatic Change, 66, 23–26. von Storch, H., Zorita, E., & Cubasch, U. (1993). Downscaling of global climate change estimates to regional scales: an application to Iberian rainfall in wintertime. J. Climate, 6, 1161–1171. Walker, G. T. (1924). Correlations in seasonal variations of weather, IX Mem. Ind. Meteorol. Dept., 24, 275–332. Wallace, J. M., & Gutzler, D. S. (1981). Teleconnections in the geopotential height field during the Northern Hemisphere winter. Mon. Wea. Rev., 109, 784–812. Wanner, H., Bro¨nnimann, S., Casty, C., Gyalistras, D., Luterbacher, J., Schmutz, C., Stephenson, D. B., & Xoplaki, E. (2001). North Atlantic Oscillation – concepts and studies. Survey in Geophys., 22, 321–381.
226 Mediterranean Climate Variability
Webster, P. J., Magana, V. O., Palmer, T. N., Shukla, J., Tomas, R. A., Yanai, M., & Yasunari, T. (1998). Monsoons: processes, predictability, and the prospects for prediction. J. Geophys. Res., 103, 14451–14510. Wibig, J. (1999). Precipitation in Europe in relation to circulation patterns at the 500 hPa level. Int. J. Climatol., 19, 253–269. Xoplaki, E. (2002), Climate variability over the Mediterranean, PhD thesis, University of Bern, Switzerland, Available through: http://sinus.unibe.ch/ klimet/docs/phd_xoplaki.pdf. Xoplaki, E., Gonzalez-Rouco, J. F., Gyalistras, D., Luterbacher, J., Rickli, R., & Wanner, H. (2003a). Interannual summer air temperature variability over Greece and its connection to the large-scale atmospheric circulation and Mediterranean SSTs 1950–1999. Clim. Dyn., 20, 537–554, doi:10.1007/s00382-002-0291-3. Xoplaki, E., Gonza´lez-Rouco, J. F., Luterbacher, J., & Wanner, H. (2003b). Mediterranean summer air temperature variability and its connection to the largescale atmospheric circulation and SSTs. Clim. Dyn., 20, 723–739. Xoplaki, E., Gonza´lez-Rouco, J. F., Luterbacher, J., & Wanner, H. (2004). Wet season mediterranean precipitation variability: influence of large-scale dynamics. Clim. Dyn., 23, 63–78. Xoplaki, E., Luterbacher, J., Burkard, R., Patrikas, I., & Maheras, P. (2000). Connection between the large-scale 500 hPa geopotential height fields and precipitation over Greece during wintertime. Clim. Res., 14, 129–146. Xoplaki, E., Maheras, P., & Luterbacher, J. (2001). Variability of climate in meridional balkans during the periods 1675–1715 and 1780–1830 and its impact on human life. Clim. Change, 48, 581–615. Yarnal, B. (1993). Synoptic Climatology in Environmental Analysis. A primer, studies in climatology series. Belhaven Press, London). Zorita, E., Kharin, V., & von Storch, H. (1992). The atmospheric circulation and sea surface temperature in the North Atlantic in winter. their interaction and relevance for Iberian rainfall. J. Climate, 5, 1097–1108. Zorita, E., & Gonza´lez-Rouco, J. F. (2000). Disagreement in North Atlantic Oscillation predictions and implications for global warming. Geophys. Res. Lett., 27, 1755–1758.
Chapter 4
Changes in the Oceanography of the Mediterranean Sea and their Link to Climate Variability Michael N. Tsimplis,1 Vassilis Zervakis,2 Simon A. Josey,1 Elissaveta L. Peneva,3 Maria Vittoria Struglia,4 Emil V. Stanev,5 Alex Theocharis,8 Piero Lionello,6 Paola Malanotte-Rizzoli,7 Vincenzo Artale,4 Elina Tragou2 and Temel Oguz9 1
James Rennell Division for Ocean Circulation and Climate, National Oceanography Centre, Southampton, UK (
[email protected],
[email protected]) 2 Department of Marine Sciences, University of the Aegean, University Hill, GR 81100 Mytilene, Greece (
[email protected],
[email protected]) 3 Department of Meteorology and Geophysics, Faculty of Physics, University of Sofia, 5, J.Bourchier Str, Sofia 1126, Bulgaria (
[email protected],
[email protected]) 4 ENEA (Ente Nazionale per le Nuove Tecnologie l’Energia l’Ambiente), Roma, Italy (
[email protected],
[email protected]) 5 ICBM, Physical Oceanography, University of Oldenburg, Postfach 2503, D-26111 Oldenburg, Germany (
[email protected]) 6 Department of Science of Materials, University of Lecce, via per Arnesano, 73100 Lecce, Italy (
[email protected]) 7 Department of Earth, Atmospheric and Planetary Sciences, Massachusetts Institute of Technology, USA (
[email protected]) 8 Institute of Oceanography, Hellenic Centre for Marine Research, Anavyssos, Greek (
[email protected]) 9 Middle East Technical University, Institute of Marine Sciences, Erdemli, Icel, Turkey (
[email protected])
4.1. Introduction In this chapter, we provide a description of the major features of the large-scale Mediterranean Sea circulation and the most important changes that have been observed in the physical oceanography of the basin. This chapter does not pretend to be a complete and comprehensive review of all the published works
228 Mediterranean Climate Variability
in the field but an introduction to the oceanography of the Mediterranean Sea aiming at an interdisciplinary audience of scientists working in various aspects of Mediterranean climate variability, the MedCLIVAR community. Most locations and topographic features mentioned in this chapter can be found in Fig. 2 of the Introduction to this book. The basic circulation of the Mediterranean Sea has long been recognized to be that of a concentration basin (Marsigli, 1681; Waitz, 1755; Nielsen, 1912) and as such its major features are quite straightforward to describe. The excess evaporation over freshwater input within the basin (Garrett et al., 1993; Gillman and Garrett, 1994) is balanced by a two-layer exchange at the Strait of Gibraltar comprising a relatively warm, fresh (15 C, 36.2 psu) upper water inflow and a relatively cool and saltier (13.5 C and 38.4 psu) outflow to the Atlantic (Bryden et al., 1994; Tsimplis and Bryden, 2000). The transformation of the inflowing Atlantic water to outflowing Mediterranean water is made through a thermohaline cell that involves the whole basin and leads to the formation of the Levantine Intermediate Water (LIW). In brief, the Atlantic inflow becomes progressively more saline as it moves eastwards. During winter it becomes cooler and denser and sinks to intermediate or sometimes deep levels. The role of the intermediate waters and in particular the LIW is important because they occupy the layer that corresponds to the sill depth (280 m), thus determining the characteristics of the Mediterranean outflow. Waters deeper than the sill depth at the Strait of Gibraltar contribute only partly to the Mediterranean outflow (Stomell et al., 1973). In addition to the formation of intermediate waters, deep water formation takes place in several parts of the basin and in particular in the Gulf of Lions (Stommel, 1972; Mertens and Schott, 1998), the Adriatic Sea (Schlitzer et al., 1991; MalanotteRizzoli et al., 1997) and recently in the Aegean Sea (Roether et al., 1996). Therefore the Mediterranean Sea acts as a reduced scale ocean as regards the thermohaline circulation and dense water formation (Bethoux, 1980; Bethoux et al., 1998, 1999). Existing climatological data sets indicate that the Mediterranean Sea is not in a steady state and is potentially very sensitive to changes in atmospheric forcing. In particular, during the last century several changes in the Mediterranean circulation have been documented. Trends in the temperature (T) and salinity (S) of the deep waters have been found in the Western Mediterranean (Lacombe et al., 1985; Charnock, 1989; Bethoux et al., 1990; Leaman and Schott, 1991; Rohling and Bryden, 1992; Bethoux and Gentili, 1999; Tsimplis and Baker, 2000) as well as the Eastern Basin (Tsimplis and Baker, 2000; Rixen et al., 2005) as well as in the upper waters (Painter and Tsimplis, 2003). The characteristics of the LIW have also been found to change in time (Astraldi et al., 1999; Brankart and Pinardi, 2001; Gasparini et al., 2005). Moreover sudden changes in the deep water
Changes in the Oceanography of the Mediterranean Sea
229
formation sites in the Eastern Mediterranean have been documented between 1987 and 1995 (Roether et al., 1996). This chapter discusses the oceanographic changes within the Mediterranean Basin concentrating mainly on the internal processes of the basin in relation to the regional forcing. Chapter 5 of this book (Artale et al.) integrates the Mediterranean Sea system to that of the North Atlantic oceanographic system and describes the characteristic mode of variability. We start by describing the forcing of the Mediterranean Sea including the atmospheric forcing, the freshwater influx from rivers, the exchange with the Atlantic Ocean and the Black Sea (Section 4.2). Then the major features of the thermohaline circulation are described (Section 4.3) before we turn to the changes that have been detected in the basin (Section 4.4) and their links to atmospheric patterns. We concentrate on the large-scale thermohaline circulation of the Mediterranean and we only examine sub-basin processes when these are directly linked with deep water formation. Thus we are not concerned with the complex combination of mesoscale and large-scale variations that dominate the surface Mediterranean circulation as seen from altimetry (Larnicol et al., 1995). However, we recognize that these also show significant variability and their contribution to the changes in the mean circulation is non-trivial (see for example, Millot, 1999). In addition to the thermohaline circulation, we discuss two other key factors that are linked to the oceanic circulation, sea level variations (Section 4.6) and the wind-wave field (Section 4.7).
4.2. The Forcing of the Mediterranean Sea We consider four forcing parameters of the Mediterranean Sea, namely, the air–sea interaction, the river influx, the exchange at the Strait of Gibraltar and the influence of the Black Sea.
4.2.1. Air–Sea Interaction The circulation of the Mediterranean Sea is determined to a large extent by the air–sea exchanges of heat and freshwater, and the wind stress forcing of the basin. The exchange of salt and water at the Strait of Gibraltar (see Sections 4.2.3 and 5.3) provides a constraint on the estimates of the basin mean values of the heat and freshwater flux. Macdonald et al. (1994) obtain estimates of the equivalent basin mean net heat loss in the range 3–7 Wm 2 using mooring-based
230 Mediterranean Climate Variability
measurements of the transport. However, the available climatological estimates of the basin mean heat flux tend to show a net heat gain which is typically in the range 20–30 Wm 2 (Bunker et al., 1982; Garrett et al., 1993). This bias is presently believed to be caused by a combination of overestimated shortwave gain arising from inadequate parameterization of attenuation due to aerosols (Tragou and Lascaratos, 2003) and water vapour (Schiano, 1996) and underestimated longwave loss (Bignami et al., 1995). Climatological annual mean fields based on the SOC climatology (Josey et al., 1999) for the net heat flux and the wind stress are shown in Fig. 73. The fields have been modified to include the Bignami et al. (1995) longwave formula as well as corrections to the shortwave formula of Gilman and Garrett (1994). A general north–south gradient in the net heat flux is apparent, from a net heat loss of up to 30 Wm 2 in the northern half of the basin to a gain of about 30 Wm 2 in the southern half. The gradient primarily reflects a reduction in the shortwave flux with increasing latitude and strong wind-driven latent heat loss in the Gulf of Lions, and the Adriatic and Aegean Seas. In winter, the heat loss in the latter three regions approaches 200 Wm–2 (e.g. Josey, 2003, Fig. 73) and is a major factor contributing to deep water formation. Significant interannual variations in the winter heat loss are known to occur, the prime example being the severe winters of the early 1990s which have been linked to the Eastern Mediterranean Transient (Theocharis et al., 1999b; Josey, 2003), discussed further in Section ‘‘when did the EMT start?’’. Broadly similar fields have been obtained in earlier studies (Bunker et al., 1982; Garrett et al., 1993).
Annual Mean Net Heat Flux and Wind Stress 45
30
-2
0.1 Nm
20 40 Latitude
10 0 -10
35
-20 -30
30 -5
0
5
10
15 20 Longitude
25
30
35
Figure 73: Climatological annual mean net heat flux (colours Wm 2) and wind stress (arrows) from the modified version of the SOC climatology discussed in the text.
Changes in the Oceanography of the Mediterranean Sea
231
The basin mean E-P may be obtained via either the terrestrial or aerological branches of the hydrological cycle (Gilman and Garrett, 1994). In the terrestrial branch, the mean E-P is determined as the sum of river runoff into the basin and the freshwater flux through the Strait of Gibraltar, while for the aerological branch, the mean E-P is equated with the divergence of the vertically integrated horizontal water vapour flux. The corresponding estimates are E-P ¼ 0.78 and 0.66 m per year from the terrestrial and aerological branches, respectively (Gilman and Garrett, 1994). The SOC flux climatology has a basin averaged of 0.71 m per year which lies between the two values noted above (Josey et al., 1999). The only region where the freshwater gains by precipitation and runoff exceed losses by evaporation is the Adriatic Sea (Raicich, 1996).
4.2.2. River Outflow Published estimates of climatological values of river discharge in the Mediterranean Sea vary significantly (Table 4) due to different methods of evaluation and/or to the different datasets used. Values for the discharge are in the range 8.1–16 103 m3 s 1 with the most recent estimates being lower than earlier estimates. Recently, Boukthir and Barnier (2000) obtain the river outflow to be 11 103 m3 s 1 while Struglia et al. (2004) obtain 8.1 103 m3 s 1 from the observations and 10.4 103 m3 s 1 by including an estimate of the errors associated with the unaccounted river basins. On the basis of the oceanic hydrological cycle, Pinardi et al. (2005) suggest that river runoff is in the range 2.2–9.8 103 m3 s 1, the upper end of this range Table 4: Published estimates of river discharge in the Mediterranean Sea. Author
Method R (103 m3 s 1)
Tixeront (1970) 16.0 Ovchinnikov (1974) 13.6 Margat (1992) 16.0 Boukthir and Barnier (2000) Struglia et al. (2004) Struglia et al. (2004)
11.0 8.1 10.4
Rain maps, data, estimate of underground waters Observations Total hydrological budget, including underground waters Observed river discharge from UNESCO dataset Observed river discharge from Med-Hycos and GRDC datasets As above, including correction to underestimates
232 Mediterranean Climate Variability
is consistent with the lower values in Table 4. The discrepancy in mean annual river discharge between earlier and more recent studies may be due to the extensive damming of the rivers, to the fact that much more water has been used in the recent years for irrigation and to the impact of changes related to the North Atlantic Oscillation which have reduced precipitation and river runoff over the Mediterranean Sea. An annual mean value for the river contribution in the Mediterranean Sea in the range 8.1–16 103 m3 s 1 (Struglia et al., 2004) makes it less than 20% of the atmospheric freshwater flux (E-P) (Mariotti et al., 2002). Geographically, the dominant contributions to the runoff are from European rivers. The discharge into the Adriatic Sea, which is the major contributor, the Gulf of Lions, and the Aegean Sea together account for 62% of the total runoff. However, the contribution of the Middle Eastern rivers is likely to be heavily underestimated (up to 60%) because there are not sufficient published data for the Turkish rivers discharging in the Levantine Basin. The North African discharge is mainly due to the Nile River which contributes about 1.4 103 m3 s 1 (Struglia et al., 2004) or as low as 540 m3 s 1 (Hamza et al., 2003) while contributions from other North African rivers are negligible (Struglia et al., 2004). The areas of maximum river discharge, namely the Adriatic, the Gulf of Lions and the Aegean Sea are also the areas where the strongest heat losses also occur in winter (Fig. 73) and are known deep water formation sites (see Section 4.3). River runoff shows a seasonal cycle with an amplitude of about 5 103 m3 s 1 the dry season being in midsummer and the peak flow in early spring (Fig. 74, from Struglia et al., 2004). The amplitude of the seasonal cycle of river outflow is almost negligible when compared to the seasonal cycle of E-P (Boukthir and Barnier, 2000; Mariotti et al., 2002). However, because the phase of the seasonal cycles of E-P and river outflow differ, the two basin-integrated components E-P and R are comparable during spring. Thus, on a sub-basin scale, river discharge variability may be responsible for some modulation of the oceanic processes (Rohling and Bryden, 1992; Zavatarelli et al., 1998; Bethoux and Gentili, 1999; Boscolo and Bryden, 2001). However, late spring freshwater discharges are unlikely to directly affect deep water formation which may probably occur earlier during the year. Observed interannual variations in runoff during the twentieth century have been up to 60% of the climatological long-term mean, while decadal variations are of the order of 20% of the long-term mean (Struglia, 2004). The interannual and/or decadal variability of river runoff is expected to be dominated by the major factor controlling precipitation over Europe, that is, the North Atlantic Oscillation (NAO). Northern European river flows are positively correlated, particularly in winter but also in spring, with the NAO Index while rivers located
Changes in the Oceanography of the Mediterranean Sea
233
Figure 74: Climatological seasonal cycle of total discharge into the Mediterranean Sea and its decomposition by continent of origin (values are in m3 s 1). From Struglia et al., 2004. in South Europe are negatively correlated (Shorthouse and Arnell, 1997). The interannual variability related to the NAO pattern has been confirmed in various Mediterranean rivers (Send et al., 1999; Cullen et al., 2002; Struglia et al., 2004; Trigo et al., 2004). The annual mean anomalies for Po, in the Adriatic, and the Rhone, in the Gulf of Lions, are shown in Fig. 75. The effects of other global patterns on the Mediterranean river discharge have not yet been resolved fully. However, ENSO has been suggested as affecting the Nile stream flow (Elfatih and Eltahir, 1996; De Putter et al., 1998) while
Figure 75: Anomalous river discharge in the Adriatic Sea (top) and in the Gulf of Lions (bottom). A reconstructed time series for the Adriatic is shown for the period 1961–84 (dashed line). Over the period 1918–96 the Adriatic time series is represented by that of the Po river (thin solid line). Discharge in the Gulf of Lions is that from the Rhone river for the period 1920–79. In both panels thick solid lines are the 5-year running means.
Changes in the Oceanography of the Mediterranean Sea
235
mid-latitude stream flow responses to extreme phases of Southern Oscillation have been detected over Turkey (Kahya and Karabo¨rk, 2001).
4.2.3. Exchanges through the Strait of Gibraltar The exchanges of the Mediterranean Sea both at the Strait of Gibraltar and at the Dardanelles are two-way flows with salinity differences driven by the water mass characteristics of the basins and mass exchanges limited by the hydraulic control points of the Straits. The exchange at the Strait of Gibraltar is of paramount importance for the Mediterranean Sea. However, as in this book it is dealt with under Section 5.3 (by Artale et al., in this book) we will deal with it briefly. Candela et al. (1989) have explained up to 80% of the observed variability of the currents in the strait on the basis of an almost barotropic mode. The inflow of Atlantic water is highly modulated by the tidal signal and the atmospheric forcing at the Strait of Gibraltar. Its barotropic component has been found to correlate well with the cross-strait sea level component (see for example, Tsimplis and Bryden, 2000). Estimates of the Atlantic inflow spanning much of the last century range between 0.72 and 1.75 Sv while outflow values range between 0.67 and 1.68 Sv (Table 5). It is not clear to what extent the variability in the estimates is a reflection of changes in the freshwater balance or an artefact of the technique used for the estimates. However omitting the oldest estimate of Scott (1915) significantly reduces the range of the inflow between 0.72 and 1.26 Sv. The net exchange is of the order of tenths of Sv (1 Sv ¼ 106 m3 s 1). The physics of the exchange are further discussed in Section 5.3 of this book.
Table 5: Published estimates of the exchange at the Strait of Gibraltar. Authors
Inflow (Sv)
Outflow (Sv)
E-P-R (Sv)
Schott (1915) Bethoux (1979) Perkins et al. (1990) Bryden and Kinder (1991) Bryden et al. (1994) Hopkins (1999) Tsimplis and Bryden (2000) Baschek et al. (2001) La Fuente et al. (2002)
1.75 1.26 1.3–1.6
1.68 1.20
0.07 0.06
0.80 0.68 0.84 0.67 0.76 0.84
0.04 0.42 0.10 0.05 0.13
0.72 1.26 0.78 0.81 0.97
236 Mediterranean Climate Variability
4.2.4. The Exchange with the Black Sea The basic dynamical feature in the Black Sea is the cyclonic rim current encompassing the sea approximately along the 1,000 m isobath and caused by the combined impacts of both mechanical and thermohaline forcing (Stanev, 1990; Oguz and Malanotte-Rizzoli, 1996; Staneva et al., 2001; Beckers et al., 2002; Kara et al., 2005). In addition, numerous mesoscale eddies created by the main flow instabilities are observed near the coast (Fig. 76). The general circulation of the BS is subject to pronounced seasonal variations as it is mainly wind-driven. Another permanent major feature of the Black Sea Basin is the Cold Intermediate Layer (CIL), which is due to the existence of a strong halocline at 70–100 m depth which limits the winter convection. The replenishment time of the CIL is estimated to be 5 years (Lee et al., 2002). As seen from Fig. 77 the amount of Cold Intermediate Water within the CIL can be substantially reduced after several warm winters thus demonstrating the sensitivity of the basin to climate variability. The Mediterranean water that enters the Black Sea sinks below the CIL as a salt wedge of 10–20 m near the Bosporus, decreasing rapidly on the shelf (<5 m) and becoming almost indistinguishable along the continental slope because of the large entrainment and dilution. The sinking is channelled by the
Sevastopol Crimea
Danube
Caucasus Kaliakra
Eas
yre ern G
tern
t Wes Bosphorus
Sakaryia
Gyr
e
Synop Kizilirmak
Batumi
Figure 76: Snapshot of sea level (cm) and surface streamlines simulated by the DieCAST model (see for more details, Staneva et al., 2001).
Changes in the Oceanography of the Mediterranean Sea
237
Figure 77: Temporal variability of the cold water content between isopycnal levels 14.4 and 15.6 sigma. Observations and simulations with the Black Sea Modular Ocean Model (Stanev et al., 2004) are shown. The solid line (right y axis) gives the winter (DJF) mean air temperature at surface estimated from ERA40 reanalysis data set. underwater extension of the Bosporus Strait and propagates further on the shallow shelf, sinking abruptly at the continental slope (Yuce, 1996; Gregg and Ozsoy, 1999; Stanev et al., 2001). Analyses of observations and numerical modelling suggest that the penetration of saline Mediterranean waters is limited to the upper 400–600 m and does not reach the bottom of the Black Sea. However, the inflowing Mediterranean water is the origin of massive intrusions of oxygen-enriched water extending to 200 km from the coast thus contributing to the lateral ventilation of the Black Sea (Konovalov et al., 2003). The inflow of water from the Black Sea to the Mediterranean is about two orders of magnitude smaller than the inflow of Atlantic water through the Strait of Gibraltar. However, the salinity difference between the Black Sea and Mediterranean Sea water is of 18 psu, that is, much larger than the 2 psu differences in the Strait of Gibraltar. Because of the large salinity differences, the role of the Black Sea outflow is not negligible, at least for the Aegean Sea. The salinity contrast between the Mediterranean and the Black Sea triggered
238 Mediterranean Climate Variability
the first ever attribution of marine circulation to density differences (Marsigli, 1681). At the Bosphorus Strait (sill depth 36 m) the Black Sea water occupies the surface layer, while the Mediterranean Sea water is observed as a salt wedge in the bottom layer. The mean position of the interface which is 10 m thick is well represented by the 20 psu isohaline. The thickness of these layers as well as their transports, are very sensitive to the sea-level difference between the basins which in turn is controlled by the basin’s water balance, regional circulation intensity and direction of wind. The two-layer density stratification separated by a transition layer is almost always observed in the Bosporus and Dardanelles Straits. However, blocking of the lower layer flow occurs during combinations of high water in the Black Sea and northerly winds (Latif et al., 1991; Yuce, 1996), which are short lived. Similarly blocking of the surface current occurs during southerly winds (Gunnerson and Ozturgut, 1974). Mixing and turbulent entrainment processes in the Straits dominate the water, salt and heat transport between the original reservoirs (Besiktepe et al., 1994; Gregg et al., 1999; Stanev et al., 2001). The mean vertically integrated transport in the Dardanelles and Bosphorus Straits on the basis of the water budget and sea level variation was estimated as 6 103 m3 s 1 (Stanev and Peneva, 2002). Recent analyses of hydro-meteorological and oceanographic data demonstrate a strong correlation between the North Atlantic Oscillation, sea level variability and thermal state of the Black Sea (Stanev et al., 2000; Stanev and Peneva, 2002; Oguz et al., 2003, Tsimplis et al., 2004). The inflowing in the Aegean Sea, Black Sea water occupies the surface layers in the North Aegean Sea where it is thought to have a controlling function on the vertical stability and mixing (Zervakis, 2000; Stanev and Peneva, 2002). Variations of the Black Sea water outflow may affect the thermohaline circulation in the North Aegean Sea (see Section ‘‘Why did the EMT happen?’’). Reductions of 100 km3/year are quite plausible, which are equivalent to changes in evaporation of 0.2 m/year over the Aegean Sea (Stanev and Peneva, 2002). By contrast, an increase of the transport of Black Sea water into the Mediterranean Sea could block or at least decrease the rates of any deep water formation taking place in the North Aegean Sea. The sea level in the Black Sea is the most important parameter controlling the exchange between the basins and it is subject to large seasonal, interannual and decadal variability. The interannual and decadal sea level variability is mainly a response to the external hydrological forcing (Stanev and Peneva, 2002; Tsimplis et al., 2004). The sea level anomalies in the time series of Batumi and Sevastopol (Fig. 78) show variations clearly related to the fresh water flux calculated as the net sum of river runoff, precipitation and evaporation over the whole Black Sea area (Simonov and Altman, 1991).
Changes in the Oceanography of the Mediterranean Sea
239
Figure 78: Annual mean sea level anomalies (mm) in Batumi and Sevastopol on the upper panel. The lower panel shows the annual mean values of fresh water flux, including river runoff, precipitation and evaporation (km3/year). For the period 1923–1997, the mean values of evaporation (E) and precipitation (P) are respectively 12 103 m3 s 1 and 7.6 103 m3 s 1 (Simonov and Altman, 1991; Stanev and Peneva, 2002). Thus evaporation exceeds precipitation but the freshwater balance becomes positive because of the river runoff. The annual discharge of the six (biggest) rivers, Danube, Dniepr, Dniestr, Southern Bug, Don and Kuban, in the Black Sea is 8.6 103 m3 s 1 (270.3 km3/ year) (Fekete et al., 1999), that is, equivalent to the river outflow in the whole of the Mediterranean Sea. There are large uncertainties in the above-stated estimates (Simmons and Gibson, 2000). Thus, further research is needed in order to understand the regional water balance and its sensitivity to climate change (Hagemann et al., 2005).
4.3. The Mediterranean Overturning Circulation 4.3.1. The Modification of the Atlantic Inflow A simplified sketch of the Mediterranean Overturning Circulation is provided in Fig. 79. The main supplier of fresh water for the Mediterranean is the Atlantic Ocean, at an average rate of about 0.8 Sv. The Atlantic inflow enters as a surface current of mainly Atlantic characteristics (S 36.5), slightly modified through mixing with the outflowing Mediterranean waters. The Modified Atlantic Water (MAW) moves towards the east following the mean cyclonic circulation
A
B
Figure 79: Sketches of the Mediterranean Overturning Circulation before the EMT (A) and during the EMT(B). Before the EMT, the MAW moved eastwards and forms about one-third of its transport as the Levantine Basin (LIW) which is then tranported westwards branching to the Adriatic and crossing back to the western basin. After crossing the Strait of Sicily it follows the coasts of the Western Mediterranean anticlockwise. There is a disagreement on whether there is a branch crossing the western basin westwards or whether large eddies (symbolized by the successive circles) are responsible for the westward transport. Deep water formation of about 0.3 Sv took place in the Adriatic Sea and the Gulf of Lions. During the EMT (B) the MAW is deflected northwards into the north of the Ionian Basin and its eastward tranport diminishes. The intermediate water formation is then deflected northwards into the Cretan Basin where about 1 Sv of deep water is formed. In this sketch, we have kept the Western Mediterranean the same as before the EMT. However evidence of impacts of the EMT in the Western Mediterranean have started being published (see Section 4.4.3 for details).
Changes in the Oceanography of the Mediterranean Sea
241
of the basin and, because of the increased warming and evaporation its salinity increases gradually to about 37.0–37.5 in the Strait of Sicily (Astraldi et al., 2002) and to around 38.6 in the vicinity of the Cretan Passages (Malanotte-Rizzoli et al., 1997). By contrast the depth of the salinity minimum increases eastward (from 20 to 100 m) during summer and autumn. However not all MAW crosses to the Eastern Mediterranean but some stays in the western basin which acts as a pool of MAW (Millot, 1999). The amount of MAW crossing to the Eastern Mediterranean is not known. However Millot (1999) indicates values around 1–3 Sv. The flow of MAW from the Strait of Gibraltar to the Strait of Sicily and then to the Eastern Mediterranean is part of the basic cyclonic circulation of the Western Mediterranean Basin which essentially is comprised of an anticlockwise cell complicated by the islands of Corsica and Sardinia. The MAW follows the North African coasts moving eastwards and forming the unstable Algerian current which generates both short-lived mesoscale eddies inconsequential for the basin-wide circulation and larger scale open sea eddies which detach from the main current and propagate seawards before turning westwards thus mixing MAW with the interior of the Western Mediterranean Basin (Millot et al., 1997; Fuda et al., 2000). At the Strait of Sicily the MAW separates into two branches, one which crosses to the Eastern Mediterranean while the other branch moves northward into the Tyrrhenian Sea flowing anticlockwise along the topography of the Italian coasts of the Western Mediterranean. At the most northern point of the Tyrrhenian Sea a further separation occurs. One branch goes north of Corsica and joins the westward flow of the second branch which first completes the circle around the Tyrrhenian Sea and then moves northwards along the Corsican and the Sardinian coasts. The rejoined branches form the Northern Current (Millot, 1999). The Northern Current continues along the French and Spanish coasts and separates into two branches one moving into the south of the Balearic Sea (Millot, 1999), while the second part continues southwards and can be found at depths greater than 200 m in the Alboran Sea as old MAW that has completed one or possibly more rounds of the Western Basin (Millot, 1999). During its circulation in the Western Mediterranean Basin the salinity of the MAW increases to 38.0–38.3 near the north coasts of the Western Mediterranean (Millot, 1999). For comparison the surface salinity at the vicinity of the Cretan passage is about 38.6 psu.
4.3.2. Intermediate Water Formation The high salinity preconditions the surface waters for intermediate or deep water formation. In the Levantine region, surface waters reach salinities of 38.9 psu.
242 Mediterranean Climate Variability
The density of the surface waters increases further through evaporation and cooling during winter until it is dense enough to sink. Depending on where the sinking water will settle, either Levantine Intermediate Water (LIW) or Levantine Deep Water (LDW) is formed. The LIW is the main water-mass of the Mediterranean Sea, occupying the intermediate layers between 200 and 500 m. Its core is usually identified by a salinity maximum of 38.95–39.05 psu near the production areas (Lascaratos, 1993) which reduces to 38.75 psu at the Strait of Sicily (Bethoux and Gentili, 1996, 1999; Astraldi et al., 1999, 2002) and to 38.4 psu at the Mediterranean outflow at the Strait of Gibraltar (Tsimplis and Bryden, 2000). The amount of LIW formed annually ranges in model outputs between 0.6 and 1.3 Sv with a typical climatological value of 1.0 Sv (Nittis and Lascaratos, 1998) which is consistent with previous estimates, based on a variety of different methods (Ovchinnikov, 1984; Lascaratos, 1993; Tziperman and Speer, 1994). One region in which either LIW or LDW has been observed to be produced is the Rhodes gyre (Ovchinnikov, 1984; Ovchinnikov and Plakhin, 1984; Malanotte-Rizzoli and Hecht, 1988; Buongiorno Nardelli and Salusti, 2000). The development of the cyclonic Rhodes gyre is crucial in the production of the LIW because of the doming of the isopycnals at the centre of the gyre (Lascaratos et al., 1993; Lascaratos and Nittis, 1998) which reduces the surface layer and brings the subsurface denser water masses closer to the influence of the surface atmospheric forcing. Modelling studies indicate that the LIW formation area can either expand over the whole of the North Levantine Basin or be shifted southwards either by changes in the mean buoyancy loss or due to synoptic-scale forcing (Nittis and Lascaratos, 1998). Other sites that have been suggested as candidates for LIW formation are the south Aegean Sea and the south Levantine Sea (Wu¨st, 1961; Bruce and Charnock, 1965; Morcos, 1972; Ozturgut, 1976; Theocharis et al., 1988; Georgopoulos et al., 1989). Intermediate waters of maximum salinity around 38.3 psu are also formed during winter in the northern coasts of the Western Mediterranean (Millot et al., 1999 and references therein). These are found underlying the MAW. They are termed the Winter Intermediate Water. Millot (1999) considers the process of formation of WIW as important in transforming the MAW into Mediterranean water types, although the exact role of the WIW is not yet resolved. The WIW follows the general circulation of the western basin underneath the MAW. The path of the LIW is normally westwards south of Crete and then filling in the intermediate waters of the Ionian Sea branching through the Strait of Otranto into the Adriatic while another branch moves towards the west possibly following the east coasts of Italy southwards and then crossing to the western basin. After it crosses the Strait of Sicily westwards, the LIW, mixes quickly and
Changes in the Oceanography of the Mediterranean Sea
243
modifies the waters from 200 to 1800 m (Millot, 1999). The intermediate waters in the western basin appear to follow a similar-to-the MAW anticlockwise pattern around the Tyrrhenian Sea and then northwards along the west coasts of Sardinia and Corsica and underneath the Northern current following the topography (Millot, 1999). However, there is significant debate on the return flow of the LIW in the Western Mediterranean. One issue is whether circulation of the LIW around the Tyrrhenian Sea actually occurs or whether the LIW crosses directly to the southern tip of Sardinia (see for example, Roussenov et al., 1995). The second issue is whether the flow of the LIW towards the Strait of Gibraltar does indeed take place following the anticlockwise circulation path suggested by Millot (1987, 1999) or whether a direct westwards flow across the western basin exists as several models indicate (Wu and Haines, 1996; Herbaut et al., 1997). A recent study (Millot and Taupier-Letage, 2005) reinforces the view of a full anticlockwise circulation and explains the LIW found in the central Algerian Basin as a result of the transport by mesoscale eddies detached from the Algerian current rather than a continuous westward flow (Millot and TaupierLetage, 2005). The disagreement between the observational analyses (Millot, 1987, 1999; Millot and Taupier-Letage, 2005) and the models which indicate the existence of a westward flow is puzzling. The authors are inclined to accept the suggestion that model deficiencies are the major cause of the discrepancy, however as the most recent data were collected in the 1990s, which is not a typical decade for the circulation of the Mediterranean due to the EMT and the associated changes in all levels of the Eastern Mediterranean, some effects in the spreading of the intermediate waters in the western basin cannot be excluded. Indeed Gasparini et al. (2005) observe increases in the density in the Sicily westward flow and a deep cascade in the Tyrrhenian deep water thus explaining the deep water trends observed by Fuda et al. (2002), Astraldi et al. (2002) in the Tyrrhenian Sea. The deeper than usual cascade of the incoming LIW could also cause an interruption of an intermediate westward flow across the western basin, although some of it would be trapped by the deep eddies detached from the Algerian current as observed by Millot and Taupier-Letage, 2005.
4.3.3. Deep Water Formation in the Eastern Mediterranean Deep water formation takes place both in the Eastern and the Western Mediterranean Basins. In the Eastern Basin, surface waters from the Levantine that have not participated in the formation of LIW continue their circulation turning northwards and westwards, to reach the northern coasts of the Aegean
244 Mediterranean Climate Variability
and Adriatic Seas. During winter under the influence of cold, dry northerly winds they become colder and even more saltier and depending on the strength of the heat losses may sink thus producing deep waters. Until recently, the Adriatic Sea was the only generally accepted area of deep water formation (Schlitzer et al., 1991; Malanotte-Rizzoli et al., 1997). The salinity of the Eastern Mediterranean Deep Water mass has always been found less than the Aegean waters outflowing the Cretan Arc Straits, thus its origin has undoubtedly been the Adriatic Deep Water (Pollak, 1951; Wu¨st, 1961; Hopkins, 1978, 1985; Schlitzer et al., 1991). These views were to change completely in 1996. The general circulation over the Southern Adriatic is permanently cyclonic and controlled by bottom topography (Manca and Giorgetti, 1998; Kovacevijc´ et al., 1999; Poulain, 2001). The intrusion of LIW through Otranto Strait (Buljan and Zore-Armanda, 1976) is a process enhancing the potential for dense water production in the Southern Adriatic (Manca et al., 2002). Winter conditions favour the intensification of the cyclonic circulation and the doming of the isopycnals in the centre of the gyre, thus facilitating open-sea convection in the Southern Adriatic. Dense water formation takes place both in the shallow Northern Adriatic and the deep southern Adriatic (Ovchinnikov et al., 1987; Manca et al., 2002). The convection process over the northern shelf has the characteristics of shelf formation, under the influence of cold and dry northerlies in the winter (Franco et al., 1982; Malanotte-Rizzoli, 1991), while the formation over the deep Southern Adriatic has the character of open ocean convection (Manca et al., 2002). The dense waters produced over the Northern Adriatic by shelf formation propagate to the south and mix with the dense water produced in the deep southern Adriatic, to form the Adriatic Deep Water (ADW) (Manca et al., 2002), which eventually overflows as an undercurrent through the Otranto Strait and constitutes the main contributor to the Eastern Mediterranean Deep Water (EMDW). The North Adriatic dense waters are characterized by densities 29.4–29.9 kg m 3 (Franco et al., 1982; Malanotte-Rizzoli, 1991). The annual export of ADW towards the Ionian Sea, measured throughout the period 1997–1999, ranged between 0.1 Sv in 1997 and 0.4 Sv in 1999, with fluctuations reaching 1 Sv (Manca et al., 2002). For the period 1985–1987, it was estimated that the renewal time for the EMDW waters formed in the Adriatic is about 126 years (Roether and Schlitzer, 1991; Schlitzer et al., 1991; Roether et al., 1994). The maximum salinity of the Adriatic appears to scale well with an index defined as the atmospheric pressure difference between the central Mediterranean and the mid-Northern Atlantic (Grbec et al., 2003). Some similarity with the low-frequency NAO variation was also evident in the same study and in Tsimplis and Rixen (2002).
Changes in the Oceanography of the Mediterranean Sea
245
The Aegean sub-basin was recognized as a potential contributor to the waters filling the deep Ionian and Levantine Basins, the Eastern Mediterranean Deep Water (EMDW) by several researchers (Nielsen, 1912; Lacombe and Tchernia, 1960). However, although dense Aegean waters have been recorded exiting the Straits of the Cretan Arc (Nielsen, 1912; Lacombe and Tchernia, 1958; Miller, 1974; El-Gindy and El-Din, 1986), their amount and density were not considered enough to constitute a significant contributor to the EMDW waters (Schott, 1915; Pollak, 1951; Wu¨st, 1961; Hopkins, 1978, 1985). It is only recently that the role of the Aegean Sea as a deep water formation area has been conclusively demonstrated (Roether et al., 1996). It is now known that the Levantine Surface Water can reach surface salinity which exceeds 39.0 in the Aegean Sea (Zervakis and Georgopoulos, 2002) and bottom water formation has been recorded over the Samothraki and Lemnos plateaux (Ovchinnikov et al., 1990; Theocharis and Georgopoulos, 1993) while the highest density ever recorded in the North Aegean is 29.64 kg m 3 (Zervakis et al., 2000). The shallow Cyclades plateau has also been suggested as a potential source of very dense waters (Theocharis et al., 1999a). Curiously the recently observed initiation of dense water formation in the Aegean Sea was accompanied by diminution of deep water formation in the Adriatic. This would imply that the production site depends on the supply of LSW and possibly LIW in order to produce deep waters and that the supply is mainly to one or the other region.
4.3.4. Deep Water Formation in the Western Mediterranean In the Western Mediterranean, deep water formation takes place in the Gulf of Lions (Stommel, 1972) where the heat-dominated local buoyancy flux appears to determine the depth of the deep water convection (Mertens and Schott, 1998). During winter, dry and cold air initially mixes the MAW and the WIW with the underlying, warmer and saltier LIW. Further heat loss leads to formation of Western Mediterranean Deep Water (WMDW). The formation of the WMDW depends on a preconditioning period, followed by violent mixing lasting for about an hour and in convection cells about 1 km in diameter. Finally, the deep water formed spreads out of the convective region (Reihn, 1995). The deep water formation appears to be a multiscale process with smallscale 1 km cells occurring within a region of about 100 km and spreading out as instability eddies of scale 5–10 km (Reihn, 1995). The rate of formation from the same event has been estimated to be between 0.2 and 0.3 Sv (Send et al., 1995, 1999) to around 3 Sv (Reihn, 1995) depending on the method used. Deep water formation does not always reach the full depth of the basin although in most
246 Mediterranean Climate Variability
years it does (Send et al., 1999). In less severe winters the waters settle at intermediate depths. Send et al. (1999) suggest that variable deep water formation driven by varying local atmospheric forcing linked to NAO-related variability is responsible for the observed changes in the deep water characteristics (see the discussion in Section 4.4.1). Significant interannual variability in the amount of the deep waters formed as well as the location is suggested by Millot (1999). In addition to the open sea formation, dense water formation may also take place on the shelf (see for example, Durrieu de Madron et al., 2005). Fuda et al. (2002) have recently suggested that deep water is also formed in the Tyrrhenian Sea although it appears that this is now resolved by Gasparini et al. (2005) as waters coming from the Eastern Mediterranean.
4.4. Climatic Changes in the Mediterranean Sea Circulation 4.4.1. Multidecadal Trends in Water Mass Characteristics Deep water changes have been observed both in the Eastern and the Western Mediterranean Basin. Some of these changes are linked, others may be independent as the sills at the Strait of Sicily restrict the exchange of deeper waters between the basins. The deep water temperature and salinity in the western basin have been found to show interannual, decadal and longer term variability. Changes in the Western Mediterranean Deep Water (WMDW) have been identified by several authors (Lacombe et al., 1985; Charnock, 1989; Bethoux et al., 1998). The WMDW has become warmer and saltier over the period 1959–1997 (Bethoux et al., 1998) with trends of 3.5 10 3 C year 1 and 1.1 10 3 year 1 for the deeper layers and trends of 6.8 10 3 C year 1 and 1.8 10 3 year 1 for the intermediate layers (Bethoux and Gentili, 1999). Various causes for these trends, including anthropogenic influence, local atmospheric conditions and hydrological conditions during dense water formation events, as well as the first signature of global warming have been proposed. Bethoux and Gentili (1999) argue that while the warming trend can be explained by greenhouse-effect-related warming of the sea surface between 1940 and 1995, the salinity trends require a rate of increase in the water deficit of the Mediterranean of the order of 0.10 m year 1. In order to achieve such a high value, it is necessary to consider not only the damming of the major Nile (Wadie, 1984) and Ebro (Martin and Milliman, 1997) rivers, but also a small increase in evaporation and decrease of the net Black Sea outflow, mainly due to Central/Eastern European river damming (Tolmazin, 1985; Bethoux and Gentili, 1996). A model study
Changes in the Oceanography of the Mediterranean Sea
247
(Skliris and Lascaratos, 2004), suggests that the river Nile damming could account for about 45% of the salinity increase in the WMDW. In examining the causes of the temperature and salinity trends of the Western Mediterranean, Krahman and Schott (1998) identified increasing trends in salinity in the surface and bottom layers, and partly attributed these trends to a gradual drop in precipitation associated with an increasing trend of the index of the North Atlantic Oscillation (Hurrell, 1995) from 1960 until 1990, as well as a decrease of the Ebro river outflow (Ibanez et al., 1996; Martin and Milliman, 1997; Tsimplis and Josey, 2001). Furthermore, Krahman and Schott (1998) related the deep water changes to local conditions in the WMDW formation area in the Northwestern Mediterranean. Estimates of trends in the Eastern Mediterranean have been included in several analyses (Tsimplis and Baker, 2000; Painter and Tsimplis, 2003; Manca et al., 2004). These analyses were made possible by the coordinated efforts that led to the development of much needed regional oceanographic databases, initially the Mediterranean Oceanographic Data Base (MODB) (Brasseur, 1995; Brassuer et al., 1996) and subsequently the MEDATLAS (Maillard et al., 2001; MEDATLAS group, 1997). Further progress was made by the inclusion of ex-USSR data (Hecht and Gertman, 2001) which cover areas in the Gulf of Libya where previously large gaps existed. Tsimplis and Baker (2000) identified significant increasing trends in the temperatures of the deep waters of the Eastern Mediterranean and some suggestions for trends in salinity, though they were reluctant to accept them as real, due to the large scattering of the salinity values especially in the Ionian Sea. Painter and Tsimplis (2003) have found that the upper waters of the entire Eastern Mediterranean have been undergoing a sustained period of cooling throughout all seasons since about 1950. The identified trends were caused by significant reductions in the winter temperatures while the other seasons did not have in general statistically significant signals. The salinity was also found to be increasing over the same period with the strongest trends often being found at the shallowest horizontal levels examined. The upper layers of the Western Mediterranean exhibit the same salinification as seen in the Eastern Mediterranean but the temperature trend is restricted to the Eastern Mediterranean. The analysis of Painter and Tsimplis (2003) supports the suggestion made by Krahmann and Schott (1998) that the LIW is not the major contributor to changes in the deep water characteristics of the Western Mediterranean implying that changes in the local atmospheric forcing are responsible for the deep water trends. However, Manca et al. (2004), have identified trends in the deep waters of the Eastern Mediterranean (see Figs. 80, 81) of similar order of magnitude as those estimated in the western basin by Bethoux et al. (1990) implying either a common atmospheric origin or a communication of the changes from one basin to the other most likely through the LIW.
248 Mediterranean Climate Variability
Figure 80: Long-term changes in (A) potential temperature ( C) and (B) salinity of the EMDW (data below 1,200 m) in the Ionian Sea Number of data points (C) in the Ionian deep waters (>1,200 m) obtained by grouping the hydrological profiles annually. The vertical bars denote one standard deviation confidence intervals. The linear regression lines and the R2 correlation coefficients are indicated. (Figure from Manca et al., 2004; reproduced with permission).
Changes in the Oceanography of the Mediterranean Sea
249
Figure 81: As in Fig. 77, for the EMDW in the Levantine Sea (also from Manca et al., 2004; reproduced with permission).
4.4.2. Rapid Changes: The Eastern Mediterranean Transient More dramatic than the observed multidecadal trends in T and S is the unexpected change in the location of the East Mediterranean deep water
250 Mediterranean Climate Variability
Figure 82: Zonal transect (middle box) of salinity by F/S Meteor south of Crete in (A) 1987 and (B) 1995. Redrawn from Roether et al., 1996.
Changes in the Oceanography of the Mediterranean Sea
251
formation that took the oceanographic community by surprise in the 1990s. The classic view of the oceanic circulation in the Mediterranean Sea had the Adriatic Sea as the major source of the deep waters (EMDW) of the Eastern Mediterranean (Pollack, 1951 and see Section 4.3.1). However, comparison of T and S sections as well as CFC-12 observations south of Crete from two F/S Meteor cruises in 1987 and 1995 (Roether et al., 1996) revealed a dramatic change of the vertical structure of the deep water column of the Eastern Mediterranean (Fig. 82). New, highly saline and rich in oxygen waters filled the bottom layers of the Ionian and Levantine Basins. These water masses were outflowing from the Straits of the Cretan Arc, ‘‘pushing’’ the older EMDW water mass to shallower depths. CFC concentrations clearly revealed the young age of these waters. Roether et al. (1996) estimated an average rate of outflow of dense Aegean waters from the Cretan Straits of about 1 Sv for seven years (1987–1995) although it is not clear from that publication how this number was derived. The new observations shook the physical oceanographic community of the Mediterranean, as they showed that significant changes in the functioning of the thermohaline circulation could occur rapidly. The Roether et al. (1996) publication was preceded by an announcement by Della Vedova et al. (1995) on changing heat transfer between the water column and the sediment, suggesting changes in the near-bottom temperature profile (also, Della Vedova et al., 2003). The evolution of this change which was termed the Eastern Mediterranean Transient was further clarified through analyses of various oceanographic cruises that took place in the region in the years between 1987 and 1993. The Cretan Sea gradually filled with newly formed, more saline water, with density exceeding 29.2 kg m 3. (Fig. 83, Theocharis et al., 1999b) at a rate reaching 3 Sv in 1991–1992. This water overflowed from the Cretan Straits filling the deep Eastern Mediterranean Basin. Estimates of the outflow based on current meter measurements covering 150 days in 1994 suggested rates close to 0.5 Sv (Tsimplis et al., 1997, 1999). Since then, the signature of the Aegean waters has propagated in the Ionian Sea (Theocharis et al., 2002; Manca et al., 2003). After the mid-nineties, the Cretan Sea returned to the pre-EMT condition of exporting small amounts of dense water that does not reach the bottom of the Ionian and Levantine Basins, but ventilates the depths of 1500–2000 m (Theocharis et al., 2002), while the main contribution of dense water for the Eastern Mediterranean is once again the Adriatic Sea (Klein et al., 2000).
4.4.3. What Caused the EMT? Major research efforts have been made during the last decade in order to understand the generation, development and demise of the EMT. Many issues
252 Mediterranean Climate Variability
Figure 83: Evolution of the pycnocline in the Cretan Sea from March 1987 to September 1997 (from Theocharis et al., 1999b; reproduced with permission). have been clarified but also alternative suggestions have been developed. We will try to explain the varying views in respect of the generation and development of the EMT in the context of four questions: When did the EMT start? Where did it start? How much deep water did it produce? and Why did it happen? The answers to these questions are interlinked and determining the location of the EMT also restricts to an extent the start time, the forcing and the quantity produced. When did the EMT Start? Several authors state that the deep water formation of 1987 was the beginning of the EMT (Klein, 1999; Lascaratos et al., 1999; Malanotte-Rizzoli et al., 1999; Theocharis et al., 1999b; Zervakis et al., 2000; Nittis et al., 2003). There is also general agreement between these authors as well as researchers looking at the atmospheric forcing (Josey, 2003; Jacobeit, 2005; Tragou et al., 2005), that the major deep water formation incident happened during the very severe winters of 1992 and 1993. However, the 1987 measurements by Roether et al. (1996)
253
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 2750 3000 3250 3500 3750
Pressure (dbar)
13
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 2750 3000 3250 3500 3750
Pressure (dbar)
Pressure (dbar)
Changes in the Oceanography of the Mediterranean Sea
A 15
17
19
21
23
25
0 250 500 750 1000 1250 1500 1750 2000 2250 2500 2750 3000 3250 3500 3750
Station No. 5 6 7 8 9 10 11 12 13 14
B
25.6
26.1
26.6
27.1
27.6
28.1
28.6
29.1
29.6
37 N
Cretan Sea
36 N
8
7
5
35 N
10
9
6
Crete
Ionian Basin
11 12
Levantine Basin
34 N
13
33 N 14
C
38.4 38.5 38.6 38.7 38.8 38.9 39.0 39.1 39.2 39.3 Salinity (psu)
D
32 N 20 E
21 E
22 E
23 E
24 E
25 E
26 E
27 E
28 E
29 E
Figure 84: The temperature (A), density (B) and salinity (C) profiles from the 1987 Meteor cruise together with the location of the stations (D). Note the salinity profiles at stations 8 and 9 in the Cretan Basin.
which are generally considered as the background condition before the EMT has started for the deep basin outside the Cretan Sea indicate that within the Cretan Sea dense waters were already forming (Fig. 84). The development of the water mass properties within the Cretan Basin can be seen in Fig. 83 (from Theocharis et al., 1999b) and Fig. 85 (from Tsimplis et al., 1999) where it is clear that the deep waters within the Cretan Basin became saltier first (between 1987 and 1991) and then (between 1991 and 1995) saltier and cooler. Thus, it appears that, as far as the Cretan Basin is concerned, the deep water formation was not an isolated incident but rather a prolonged one with at least two steps. The two-step scenario has also been suggested by Lascaratos et al. (1999) and Theocharis et al. (1999b). The first step involves increase in salinity and the second one is associated with cooling and further salinity increase. The salinity profile at station 9 (Fig. 84) from the 1987 Meteor cruise indicates that saltier waters were already forming near the continental
254 Mediterranean Climate Variability
14.60 M10 M8 O58 O62 P45 P76 L44 L46
14.52
Potential Temperature (oC)
14.44 14.36
} }
}
1987
1991
1991 1995 1987
14.28 14.20 14.12 14.04 13.96
1995 13.88 13.80 38.6
A 38.8
38.9
39.0
39.2
39.3
Salinity
O62
O58 L44
P45
M8 L46 P76 M10
B
Crete
Figure 85: Temperature and salinity plots at selected stations showing the stages of development of the EMT (Tsimplis et al., 1999).
Changes in the Oceanography of the Mediterranean Sea
255
shelf of Crete at that time. These observations suggest a number of possibilities: (1) Deep water formation was not the exception but the norm during, at least the winters of the period 1987–1995 although in some years they may have resulted into smaller volumes of deep water which remained undetected. (2) The 1992–93 deep water formation was linked to the atmospheric forcing identified by Theocharis et al. (1999b) and Josey (2003), but deep water formation had started earlier, by other processes affecting mainly the salinity of the upper waters. These could be linked to river damming (Rohling and Bryden, 1992; Boscolo and Bryden, 2001) or to the continuous reduction of freshwater at the Eastern Mediterranean through the shift of the NAO to an almost permanent high state during the last few decades (see for example, Tsimplis and Josey, 2001) or due to reduction of freshwater in the Northern Aegean because of reduced outflow from the Black Sea (Zervakis et al., 2000). If only the surface salinity is increased then the maximum density will be formed at higher temperature which is what we observe in Fig. 85 in agreement with Tsimplis et al. (1999). Thus we think that long-term processes contributed in the preconditioning of the EMT and the water mass characteristics of the deep water formed. Where did the Deep Water Formation Take Place? This is clearly a surprise question for non-experts in Mediterranean Sea oceanography. If the above-made suggestion that the long-term changes in freshwater fluxes contributed to the EMT is correct, then the spatial distribution of such changes must, of course, be distributed to the Eastern Basin through the oceanic circulation. Thus, it is in respect of the events triggering the massive deep water formation that the views in respect of the position of the EMT differ. Clarifying the location is, of course, tied to the identification of the cause and understanding the sensitivity of the basin to local changes in the freshwater and heat forcing. The first studies of the EMT suggested the Cretan Sea as the sole area of deep water formation (see for example, Roether et al., 1996; Klein et al., 1999; Lascaratos et al., 1999; Malanotte-Rizzoli et al., 1999; Theocharis et al., 1999b) arguably because this was the place where the changes in the deep water characteristics were first observed. Accordingly, most modelling studies have concentrated on exploring the contribution of the atmospheric forcing on the Cretan Sea and its circulation changes. These studies developed the dominant school of thought which suggests that changes in the wind field caused diversification of the MAW to the north of the Ionian Sea (Malanotte-Rizzoli et al., 1999) thus depriving the Levantine from fresher water and led to the generation of the EMT through several circulation changes discussed in Section ‘‘Why did the EMT happen?’’ (Malanotte-Rizzoli et al., 1999). According to this view, the Cretan Sea is the place of deep water formation during the EMT.
256 Mediterranean Climate Variability
The suggested alternative mechanism was a reduction of the Black Sea freshwater into the Northern Aegean (Zervakis et al., 2000). Normally, this water mass being less saline than the underlying water masses acts as a lid that absorbs atmospheric forcing thus not permitting the underlying denser water to be exposed to the highest heat losses directly and thus form deep water. However, Zervakis et al. (2000) observed that the published minima in the Black Sea outflow coincide with the deep water formation periods and are also supported by the available oceanographic observations in the North Aegean. Thus, they suggest that deep water formation started in the North Aegean Basin and then overflowed over the Cyclades Plateau into the Cretan Sea and contributed in the preconditioning for the 1992–93 event. It is worth noting that the deep basins of the North Aegean underwent full replenishment of their waters in March 1987. It is of course possible that both the above mechanisms were triggered in parallel but clearly the relevant importance of each of them is crucial in understanding the system. How Much Deep Water was Produced by the EMT? The characterization of the EMT is mainly linked to the outflow of waters over the sills of the Cretan Straits. The initiation of outflow requires production of deep water in quantities capable of filling the Cretan Basin to sill level and the outflowing to the deep basins of the Eastern Mediterranean. As an alternative, one can consider that small amounts of deep waters were formed year after year either until the dense waters of the Cretan Basin reached the sill level and then a much smaller deep water production initiated the outflow. Roether et al. (1996) have estimated the EMT outflow as 1 Sv over the period 1986–1997 but without providing a description of the estimation or any error bars. Theocharis et al. (1999) suggest that the spring and summer outflow must be stronger than the outflow in autumn and winter as deep water formation is a late winter event. Moreover, it is also reasonable to assume that the flow must be reducing as the head of the denser water is reduced. Thus, during the spring following the deep water formation the outflows may have been higher than 1 Sv. However, estimates of the outflow from subsequent current meter measurements suggest lower overall values. A simple hydraulic model estimating the flow by reference to the inclination of isopycnals over the sills (Tsimplis et al., 1999) give values about half those given by Roether et al. (1996). Consistency between the current meter measurements and the estimate of Roether et al. (1996) can be achieved if one accepts that the suggested higher outflows during spring time happened only during the years of deep water formation, namely during 1987, 1992 and 1993 (Theocharis et al., 1999). If 1 Sv is the average over 7 years and we accept that in total the observed values from the various current meter measurement (Tsimplis et al., 1999) indicate an average of 0.5 Sv during the four non-formation years as
Changes in the Oceanography of the Mediterranean Sea
257
well as the autumn and winter of the three formation years then the outflows following the formation periods must be in excess of 2.5 Sv in support of the suggested outflow rates of around 3 Sv during the formation period (Theocharis et al., 1999). However, if formation started in the North Aegean and filled in that basin first before outflowing over the Cyclades plateau to the Cretan Sea (Zervakis et al., 2000) then arguably even higher rates could be suggested. Why did the EMT Happen? We have already discussed several of the scenarios suggested as contributing to the generation of the EMT, namely long-term forcing through freshwater reduction from rivers, the NAO pattern or the Black Sea outflow, anomalous heat fluxes during specific winters, specific changes in the circulation of the MAW and the LIW. We now re-examine them and link them to evidence from modelling and observational studies. Modelling studies have suggested competing or contributing mechanisms and as they cannot all be correct they cannot be conclusive as to how the EMT has been produced. However, they provide valuable insights on the potential contribution of each mechanism separately and the size of the required changes. Local (over the southern Aegean) anomalous meteorological conditions may have been partly responsible for the shift of dense-water formation activity from the Adriatic to the Aegean Sea (Lascaratos et al., 1999 supported by observations by Theocharis et al., 1999b). Samuel et al. (1999) showed that the period of increased dense-water formation over the Aegean Sea coincided with change in the wind-driven circulation of the Eastern Mediterranean which supplied more LIW to the Aegean Sea (Fig. 86). Malanotte-Rizzoli et al. (1999) identify significant changes in the oceanic circulation: (1) a gyre developed in the Ionian Sea which deflected the fresher Atlantic Water from its course towards the Levantine and the Aegean Sea resulting in increased salinities in both basins, (2) gyres developed east of the eastern Straits of Crete that blocked the normal progress of the LIW south of Crete and forced the LIW into moving through the eastern Cretan Straits into the Cretan Sea thus increasing the local salinity, (3) this situation which was observed both in 1991 as well as 1995 (LIWEX Group, 2003) also involved Cretan Intermediate Water (CIW) spreading out from the Cretan Sea towards the Sicily Strait. These changes have been linked to changes in the atmospheric forcing (Pinardi et al., 1997). Wu et al. (2000), in a model study, have shown that a 2 C SST anomaly for seven years over the North Aegean could generate the outflow of large quantities of dense waters through the Cretan Straits. Such changes have been further supported by Zervakis et al. (2000) who identified two major dense-water formation
258 Mediterranean Climate Variability
A
B
Figure 86: Wind stress from the climatology of the Southampton Oceanography Centre for the periods (A) 1980–1987, and (B) 1988–1993. From Samuel et al., 1999 reproduced with permission. events in the North Aegean, in 1987 and 1992–93. They proceeded to suggest that the 1987 event was the trigger of the EMT, by accelerating the Aegean circulation ‘‘pulling’’ highly saline waters from the Levantine northwards, thus preconditioning for the second, larger formation event in 1992–93 (Zervakis et al., 2004). In assessing the factors causing the 1987 formation, they suggested that reduced Black Sea outflow could help the erosion of the North Aegean pycnocline. Josey (2003) analysed heat and buoyancy flux data from ECMWF reanalyses and
Changes in the Oceanography of the Mediterranean Sea
259
identified anomalously high buoyancy losses from the Aegean in the winters of 1992–93. However, his analysis does not reveal any anomalous air–sea interaction in winter 1987 nor in the rest of the period 1987–1995. Rupolo et al. (2003), forcing a Mediterranean OGCM with daily satellite SST and ECMWF wind stress for the years 1988–1993, have reproduced the main characteristics of the EMT showing that the phenomenology of this event can be separated in two distinct periods as suggested in Section ‘‘When did the EMT start?’’ During 1988–1991, a wind-induced modification of intermediate layer circulation caused an increase in salinity in the Aegean Sea and a contemporary decrease in the inflowing intermediate water in the Adriatic. These changes in the intermediate circulation make the Aegean Sea the favourite site for the production of dense water and the EMT fully develops during the cold winters 1992 and 1993 when the dense water of Aegean origin fill the deep layers of the Ionian Sea, substituting the ‘‘old’’ ADW. Recently, data provided by ex-USSR cruises in the Aegean suggest that the pre-1990 deep-water formations took place north of the Cyclades (Gertman et al., 2005) and support the Malanotte-Rizzoli et al. (1999) scenario of salinification of the Aegean through blocking of the MAW entering the Levantine Sea by the development of a gyre in the north Ionian Sea. Boscolo and Bryden (2001) suggest that the damming of the Nile and the ex-USSR rivers flowing into the Black Sea lead to an effective increase of the freshwater deficit over the Eastern Mediterranean. This process could erode the stratification in the Cretan Sea and produce dense water thereby initiating the EMT. However, river outflow measurements do not support long-term reduction in the river outflow of the Black Sea (Stanev and Peneva, 2002). On the contrary, observed reduction of the evaporation is likely to have led to increased freshwater outflow in the Aegean Sea (Tsimplis et al., 2004). Skliris and Lascaratos (2004) based on a model analysis, find that a significant part of the observed salinity changes could be due to the effect of damming of the Nile thus implying significant preconditioning due to this anthropogenic cause. However, they also note that the effect of damming is probably smaller than the effect of long-term changes in the E-P balance (Skliris and Lascaratos, 2004). Tsimplis and Josey (2001) suggest that the NAO-induced changes in oceanic E-P as well as river outflow is larger than the suggested freshwater input reduction cause by the damming of the Nile. The relative importance of long-term slow changes like the damming of rivers and the NAO index, against seasonal processes like abnormally cold winters, and short timescale processes such as extremely cold days associated with relatively rare synoptic features remains an open issue in the debate of the EMT. Similarly, it is unclear whether localized changes like a sudden reduction of the Black Sea outflow, or deprivation of the Eastern Mediterranean from freshwater through the damming of the Nile are more important than the basin overall water
260 Mediterranean Climate Variability
and salt budget. Whether the forcing was caused by blocking of the Atlantic inflow or was due to local (in the southern Aegean and Levantine) processes or to processes in the North Aegean needs also to be clarified. Thus, the studies of the EMT challenge our knowledge about both the temporal and spatial scales involved and our ability to separate and quantify the contribution of the various forcing mechanisms.
4.5. The Impact of Large-Scale Atmospheric Variability on the Mediterranean Sea Large-scale patterns of atmospheric variability, for example the North Atlantic Oscillation (NAO), that have their primary centres of action over neighbouring ocean basins have the potential to exert an influence on the Mediterranean Sea. It is well known that the NAO has a significant impact on Mediterranean climate, in particular the amount of precipitation (e.g. Hurrell, 1995) the river runoff (Struglia et al., 2004) and the sea level variability (Tsimplis and Josey, 2001). Moreover, a sudden change observed in the LIW characteristics over the winters of 1981 and 1982 (Brankart and Pinardi, 2001) was attributed to surface heat budget changes over the basin. Rixen et al. (2005) suggest an NAO influence on the Western Mediterranean Deep Water similar to that effected by the NAO in the North Atlantic. They also suggest that the anticorrelations found between the Mediterranean SSTs and the NAO as well as the anticorrelation found by Tsimplis and Rixen (2002) between the upper water temperatures in the Adriatic and the Aegean Seas and the NAO has probably played an important role in establishing the Eastern Mediterranean Transient (Demirov and Pinardi, 2002). However, the extent to which large-scale patterns play a role in the variability of the ocean circulation, in particular the major EMT event discussed in Section 4.4 is not well established. Some progress has been made and the link between surface forcing and the EMT, together with its possible relationship to large-scale climate patterns has been discussed earlier. Tsimplis and Josey (2001) have suggested a link between the long-term changes in the deep water of the Mediterranean, the observed sea level changes and the development of the EMT with the consistently high values of the NAO index during that period. Although they were justified in their assessment about the sea level changes and their assertion about the NAO effect in respect of the deep water changes is also shared by other researchers (see for example, Millot, 1999; Send, 1999) their suggestion about the link between the NAO and the EMT appears less strong. The latter suggestion was based on their observation that between 1988–93, the freshwater changes due to changes in the E-P which peaked over the
Changes in the Oceanography of the Mediterranean Sea
261
Aegean and changes in the river outflow were sufficient to cause, at least partly the observed freshwater deficit that has been suggested as a cause of the EMT. However, subsequent work by Josey (2003) suggests that the E-P contribution was probably not significant while Tsimplis et al. (2004) estimate, on the basis of an analysis of sea level variations in the Black Sea an increasing outflow of Black Sea water due to local reduction in evaporation over the Black Sea. Thus, it presently appears that any contribution from the NAO is likely to be affecting the long-term preconditioning of the waters only (as suggested by Tsimplis and Rixen, 2002 and Rixen et al., 2005) but it did not trigger the deep water formation incidents of 1991–92 and 1992–93. In contrast, Josey (2003) suggests that the very cold winters during which the EMT took place are linked to anomalously high pressure over Western Europe and the North-East Atlantic thus implying another major mode of atmospheric variability which is potentially related the East Atlantic Pattern (EAP). This suggestion is supported also by Dunkeloh and Jacobeit (2003) and Jacobeit (2005) who identified a significnat change in the second canonical correlation pattern of geopotential height and precipitation after 1987 (Fig. 74 from Dunkeloh and Jacobeit, 2003). However, it is clear that the EMT has started before 1991 thus other factors must have contributed to its initiation as there were no significant thermal or haline forcing anomalies prior to this time (Josey, 2003). The extent to which these factors are linked with large-scale atmospheric patterns or were just coincidental is not presently resolved. Is the EMT a unique phenomenon or has it been happening in the past without us noticing? Some suggestions exist for an earlier EMT in the very noisy hydrographic data of the Ionian Sea. Moreover, extreme winter heat loss also occurred in the mid-1970s and this could have been expected to produce conditions conducive to deep water formation in the Aegean Sea (Josey, 2003). However, the significance of this earlier extreme forcing is complicated by longterm river runoff trends (Rohling and Bryden, 1992) which may have led to stronger stratification in the Aegean Sea and greater resistance to convection at this time. Thus it presently remains unclear whether an event similar to the EMT occurred during the mid-1970s.
4.6. Sea Level Changes in the Mediterranean Sea The Mediterranean sea level depends on the pressure gradients across the Strait of Gibraltar, the prevailing hydraulic conditions at the Strait (Ross et al., 2000), the steric variations and changes in the water budget driven by the regional atmospheric forcing (Tsimplis and Josey, 2001) and the circulation within the basin. A few long sea level records spanning to the beginning of the 1900s exist
262 Mediterranean Climate Variability
in the Mediterranean Sea and these are located at the northern coasts of the Western Mediterranean (Marseille and Genoa) and at the northern coasts of the Adriatic Sea (Trieste and Venice) (Tsimplis and Spencer, 1997; Tsimplis and Baker, 2000). The sea level trends for the three longer stations are in the range 1.1–1.3 mm/year, that is, slightly lower than the estimated global value for sea level rise which is in the range of 1–2 mm/year (Church et al., 2001). A fourth long-term tide gauge record in Venice has particularities due to variations in local subsidence and therefore it is not usable in the context of sea level trends during the period when ground water extraction cause local subsidence (Woodworth, 2003). New data for Trieste extending the record for this tide gauge fifteen years back to 1875 leave the above estimates unaffected (Fabio Raicich, personal communication). The Mediterranean sea level variability can be separated into three periods. Until the 1960s, sea level in the Mediterranean Sea had trends equivalent to those at the open ocean stations (Tsimplis and Baker, 2000). During the second period, between 1960 and the beginning of the 1990s sea level in the Mediterranean Sea was either not changing or decreasing (Orlic and Pasaric, 2000; Tsimplis and Baker, 2000) mainly due to atmospheric pressure changes during the winter period (Tsimplis and Josey, 2002; Woolf et al., 2003) as well as temperature (T) reduction and salinity (S) changes linked to the NAO (Tsimplis and Rixen, 2002). It appears that these T and S changes are restricted to the northern part of the basin, namely the Aegean and Adriatic Sea (Painter and Tsimplis, 2003) while there are basin-wide east–west gradients on atmospheric pressure as well as E-P (Tsimplis and Josey, 2001). The third period of interest between 1993 and 2002 is based on analyses of the TOPEX/POSEIDON dataset (Cazenave et al., 2001; Fenoglio-Marc, 2002) which reveal a picture much more complicated than that of a coherently varying basin. During this period fast sea level rise was observed at the Eastern Mediterranean Sea (Cazenave et al., 2001; Fenoglio-Marc, 2002) and was linked with changes in observed sea surface temperature (Cazenave et al., 2001). Recently, on the basis of altimetric data, Vigo et al. (2005) confirmed an abrupt reduction of sea level rise rates (Fenoglio-Marc, 2002) as well as negative trends in parts of the Eastern Mediterranean Sea after 1999. These changes appear consistent with sea surface temperature changes and it is suggested by Vigo et al. (2005) as being a consequence of the restoration of the Adriatic Sea as the main source of deep water in the Eastern Basin. The observed sea level values and their temperature-related forcing has been confirmed by the use of climatological data of oceanic temperatures (Tsimplis and Rixen, 2002) for part of the 1990s. During the same period of time a reduction in the sea level gradient across the Strait of Gibraltar has been observed and the change has been suggested as caused by varied hydraulic conditions in the Strait (Ross et al., 2000) or by
Changes in the Oceanography of the Mediterranean Sea
263
changes in the density difference between the Mediterranean and the Atlantic (Brandt et al., 2004). However, a reverse jump in the sea level slope across the Strait occurs in 1999 (Tsimplis et al., 2005b) in the slope and this is accompanied with the sea level rise reduction or reversal observed in the Eastern Mediterranean Sea (Fenoglio-Marc, 2002; Vigo et al., 2005). Thus it is possible that there is a link between the Eastern Mediterranean sea levels and the hydraulic jumps in the Strait of Gibraltar, possibly through the salinity changes in the LIW (Tsimplis et al., 2005b). Between 1958 and 2001, the tide-gauge records indicate sea level trends in the range of 0.8–0.4 mm/year 0.4 mm/year (Tsimplis et al., 2005a). During the same period, the direct meteorological forcing caused sea level reduction of 0.4 to 0.6 mm/year linked with the North Atlantic Oscillation and particularly increasing atmospheric pressure. After the removal of the meteorological influence from the sea level records the resulting trends were found to be 0.3 0.4 mm/year at the Western Mediterranean and 1.3 0.4 mm/year at the Eastern Mediterranean which is strongly affected by rapid sea level rise in the period 1993–2001 with rates of 5–10 mm/year probably related to the Eastern Mediterranean Transient (EMT) (Tsimplis et al., 2005a). In conclusion, during the last 40–50 years sea level trends within the Mediterranean Basin differ significantly from those of the nearby Atlantic Ocean (Tsimplis and Baker, 2000; Woolf et al., 2003). It is yet unclear for how long can the Mediterranean Sea sustain sea level behaviour different from the open ocean. To the extent that the differences are caused either by largescale meteorological patterns like the NAO or basin or sub-basin steric processes which do not affect the pressure gradient across the Strait of Gibraltar, substantial differences may be permitted to remain. However, if a mechanism of mass addition to the open ocean by melting ice is assumed as the primary cause of sea level rise as suggested by Miller and Douglas (2004) and provided that such a mechanism is enhanced with time (Church et al., 2001) it is unlikely that the Mediterranean Sea will be able to sustain its distinctive behaviour for more than 20–30 years.
4.6.1. Extreme Sea Levels Changes in extreme sea levels are arguably more important than mean sea level rise as they pose significantly higher risks to coastal regions. Within the Mediterranean Sea very few studies on extreme sea levels have been conducted and even fewer are concerned with changes in extremes. Moreover, the published studies are not basin-wide but rather regional and limited in scope. Lionello (2005) has analysed the trends of extremes storm surges on the basis of
264 Mediterranean Climate Variability
the records in Venice and found no significant trend since 1940 apart from that produced by the combination of sea level rise and local ground subsidence. Raicich (2003) has investigated sea level extremes in Trieste for the periods 1939–2001 and found a decreasing trend in strong positive surges in spite of increases in southerly winds due to increased atmospheric pressure (Raicich, 2003; see also Pirazzoli and Tomasin, 2002, 2003; Trigo and Davies, 2002). Tsimplis and Blackman (1997) have documented the sea level extremes for the Aegean Sea for a period of eight years (1982–1989) but no information on trends of extremes could have been derived with these short time-series. However, Tsimplis and Blackman (1997) suggest that the observed extremes are in most cases common in the whole of the Aegean Basin and are consistent with a linear addition of the extreme pressure and wind effects thus implying that knowledge of these fields would suffice for estimating changes in the sea level extremes in the Aegean Sea. A global study (Woodworth and Blackman, 2004) suggests that extreme sea levels have been increasing during the last decades at the European coasts and that these changes are consistent with the impacts of the North Atlantic Oscillation. This study had been heavily influenced by the northern European stations and thus its results are not representative of extreme sea levels within the Mediterranean Sea. The mean sea level changes described above would cause extreme sea levels to decline between the 1960s and the 1993 and to increase afterwards. As the NAO remains the major atmospheric forcing pattern during winters until 2001 (Tsimplis et al., 2005a) it is unlikely that significant changes in the maximum surges would have occurred and the changes in the mean sea level which was NAO controlled at the Eastern Mediterranean until 1993 and EMT controlled after 1993 are likely to be the dominant change in the sea level extremes. The problem of extremes in the Mediterranean Sea is complex. In this book, it is also discussed in Chapter 6, in relation with the cyclones in the Mediterranean region. It is worth noting that floods have been observed with non-exceptional atmospheric conditions in the east Adriatic when resonance between travelling air pressure disturbance with a coastal wave (Vilibic et al., 2004). Moreover in the Adriatic in addition to the contribution of storm surge, seiches and tides, low-frequency oscillations (0.01
Changes in the Oceanography of the Mediterranean Sea
265
in other European stations during the second half of the last century (P. Pirazzoli, personal communication).
4.7. Changes in the Wind-wave Field Studies on the changes in the wind-wave fields are hindered by the lack of data. In particular, wave buoy observations are only available in limited locations and only since the early 1990s. Non-directional satellite observations of significant wave height are also available for 1993–2002 as two measurements per day above the Mediterranean Sea. Thus the major source of wave-field studies is model simulations based on the atmospheric forcing derived from reanalyses of atmospheric data. However, such simulations because of their rough resolution are affected by systematic surface wind under-evaluation. This problem affects both the analysis of monthly average SWH (Significant Wave Height) fields and extremes SWH values, which are presented in Section 6.3.4. The variability of the monthly average SWH (Significant Wave Height) field in the Mediterranean Sea, in the period 1958–2001 has been analysed (Lionello and Sanna, 2005) using the data provided by simulations carried out using the WAM model (WAve Model cycle 4) forced by the wind fields of the ERA-40 (ECMWF Re-Analysis). Comparison with buoy observations, satellite data, and simulations forced by higher resolution wind fields indicate that, apart from the underestimation of the observed SWH, space and time variability of the wave fields are correctly simulated by the models. The SWH field shows large inter-annual and inter-decadal variability and a statistically significant decreasing trend in mean winter values of ( 0.2 cm/year) mainly caused by a weakening of the Mistral wind regime during winter (Fig. 87). When the mean annual SWH values are considered a much smaller trend of 0.08 cm/year is found. The correlation coefficient between the winter average SWH and the winter NAO index is 0.47 and 0.73 for the seasonal and 5-year low-pass filtered time series, respectively (Fig. 88). In spite of the significant correlation, the link between Mediterranean SWH variability patterns and NAO is not that strong. The SLP composite, based on winter months of the average SWH shows a pattern different from the NAO (Fig. 89, upper panel). This composite has been obtained subtracting the average of the set of fields when the average winter SWH was in the 10% lowermost range to that when it was in the 10% uppermost range. The composite is not similar to the NAO dipole. It presents a tri-pole with a principal minimum located over central Europe, a minor one in the Atlantic and a large maximum over North Atlantic, extending to Russia. On the Mediterranean
266 Mediterranean Climate Variability
Figure 87: Time series of the average winter SWH (values in meters), linear regression (straight line) and 95% confidence interval (dashed line) (from Lionello and Sanna, 2005). region, it determines a strong atmospheric circulation, following the shape of the basin, with a prevailing south-eastward direction in the western basin and north-eastward in the eastern one. During summer, the wave field variability appears weakly correlated with the Indian Monsoon index (Figs. 88, 89, lower panels) with correlation coefficients 0.34 for the summer values and 0.5, which is not significant for the 5-year low-pass filtered time series. This correlation reflects the moderate influence of the Monsoon on the meridional Mediterranean circulation. Thus the SLP patterns associated with the SWH inter-annual variability reveal structures different from the NAO and Monsoon circulation although these indices are significantly correlated with the winter and summer SWH, respectively. This is because the wave-field variability is conditioned by regional storminess in combination with the effect of fetch. The latter is likely to be the most important. Thus, although the role of the large-scale patterns (mainly NAO) influences the average SWH field, their effect is strongly modulated by mesoscale factors and geographical land–sea distribution. The fetch acts as a filter, selecting surface atmospheric circulation components where regional characteristics
Changes in the Oceanography of the Mediterranean Sea
267
Figure 88: Time series of winter NAO index (dashed black line) and winter SWH index (light grey continuous line) (top). Time series of Indian Monsoon index (dashed black line) and winter SWH index (light grey continuous line) (bottom). A 5-year low-pass filter has been applied to all time series (from Lionello and Sanna, 2005).
conform to the shape of the basin and, therefore, are more effective in producing waves. This modulation implies that variability regimes reflect regional features, directly responsible for the wave generation, more than large-scale patterns, though a link to NAO and Indian monsoon is certainly present.
268 Mediterranean Climate Variability
Figure 89: SLP composites based on monthly average SWH deviation from annual cycle for winter (top) and summer (bottom). Contour levels according to the level bars; values in hPa. These composites are obtained subtracting the average of the set of fields when the average SWH was in the 10% lowermost range to that when it was in the 10% uppermost range. Contour levels according to the level bars; values in hPa (from Lionello and Sanna, 2005).
4.8. Outlook and Future Challenges During the last decade the Mediterranean Sea has been an exceptional place to do oceanographic and climatic research. The slow changes in the water mass characteristics of the deep waters of the Mediterranean Sea and their potential links to atmospheric forcing and/or to damming of the rivers together with the rapid changes in the deep water formation locations as evidenced
Changes in the Oceanography of the Mediterranean Sea
269
by the EMT and the uncertainty that exists both on the exact forcing mechanisms as well as the possibility of being a regular pattern which we managed to observe for the first time have raised issues about the coupling between the rapid and the longer term changes, their respective significance and the triggering mechanisms. Moreover, the exceptional character of observed sea level changes which appear not linked with the global sea level situation raise the question whether the Mediterranean Basin future sea level scenarios cannot be based on the global ones as these do not include the relevant forcing mechanisms. Have these changes affected the Mediterranean society? Estimating the cost of economic or social activities related to the use of the sea by the Mediterranean countries is not easy. Rough estimates of the value of the economic activities by the first author suggest that the total of the sea-related economic activities does not exceed 5% of the GDP for each country. In addition to the economic implications for society, ecosystem changes may become important. The ecosystem of the Mediterranean and the Black Seas has been characterized as ‘‘sensitive’’ and the future not ‘‘rosy’’ (Turley, 1999). The major influence appears to be linked with eutrophication caused by increased agricultural phosphates and the damming of rivers (Turley, 1999) rather than to climatic changes. Dust deposition which carries Fe and is very important for phytoplankton growth is, according to Turley (1999), controlled by the NAO. Consequently, the varying dust input is likely to affect climate change as well as the availability of nutrients in the basin although the results are presently unpredictable (Turley, 1999). The multiscale variability of the Mediterranean Sea has serious implications for policy making and environmental management (Zavatarelli, 1999). It is of the highest priority for the scientific community to assess the impact the observed oceanographic changes have on the Mediterranean society and prioritize in accordance with these impacts. Preliminary assessments of some specific processes like the effect of climate change on estuaries have been published for particular countries (Paskoff, 2004) and national impact assessments are presently underway is several countries (Nadia Pinardi and Enrique Alvarez Fanjul, personal communications). It is therefore necessary to develop a Mediterranean Impact Assessment study exploring the vulnerability of the Mediterranean Sea natural and socio-economic system. This would require interaction with social scientists and engineers as well as with governing bodies which will need to assess the efficiency of their decisionmaking mechanisms. Whether this is a feasible process for all Mediterranean countries is unfortunately doubtful due to the regional geopolitical complexities. Thus it is important that the developed Mediterranean countries support the developing ones in establishing monitoring systems at their coasts as well as performing their national impact assessments.
270 Mediterranean Climate Variability
The developments in satellite oceanography has led to the realization that smaller scale processes are present and need to be accounted for in order to understand changes in the system. Thus present efforts of long-term monitoring should be extended and improved (CIESM, 2002). The identification of the main modes of temporal and spatial variability of circulation in the Mediterranean Sea and their response to different modes of atmospheric variability are not yet fully resolved. Within this context the analysis the relationship between changes in the Mediterranean circulation, deep water formation, sea level and regional climate variability on the one hand and local synoptic scale processes on the other hand should be promoted. Exploratory assessments of future changes in Mediterranean circulation, water mass characteristics and sea level under scenarios of a warming atmosphere would be useful tools both scientifically and for policy-making reasons. Similarly for the Black Sea there are still many issues to be addressed. The research is to be directed towards interdisciplinary issues. Coupled models (atmosphere–ocean, or physical–biogeochemical models) have to address the various feedbacks between the components of the climate and biogeochemical systems. For the numerical modellers the challenge should be to unify the adjacent basins (Mediterranean, Marmara, Black and Azov Seas) addressing the climatic controls of straits.
Acknowledgements We would like to thank Beniamino Bruno Manca and Salvatore Marullo for kindly providing original figures. The first author was partly supported from the Tyndall Centre for Climate Change Research (UK). We would also like to thank one anonymous reviewer for very constructive comments on an earlier draft of this chapter.
References Artale V., Calmanti, S., Malanotte-Rizzoli, P., Pisacane, G., Rupolo, V., & Tsimplis, M., The Atlantic and Mediterranean Sea as connected system, Chapter 5 (this book). Astraldi, M., Balopoulos, S., Candela, J., Font, J., Gacic, M., Gasparini, G. P., Manca, B., Theocharis, A., & Tintore´, J. (1999). The role of straits and channels in understanding the characteristics of Mediterranean circulation. Progress in Oceanography, 44(1–3), 65–108. Astraldi, M., Conversano, F., Civitarese, G., Gasparini, G. P., Ribera D’alcala´, M., & Vetrano, A. (2002). Water mass properties. Journal of Marine Systems, 33–34, 155–177.
Changes in the Oceanography of the Mediterranean Sea
271
Baschek, B., Send, U., Lafuente, J. G., & Candela, J. (2001). Transport estimates in the Strait of Gibraltar with a tidal inverse model. J. Geophys. Res., 106(C12), 31033–31044, 10.1029/2000JC000458. Beckers, J.-M., Gregoire, M. L., Nihoul, J. C. J., Stanev, E., Staneva, J., & Lancelot, C. (2002). Modelling the Danube-influenced North-Western continental shelf of the Black Sea. I: hydrodynamical processes simulated by 3-D and box models. Estuarine Coastal and Shelf Science, 54, 453–472. Besiktepe, S., Sur, H. I., Ozsoy, E., Latif, M. A., Oguz, T., & Unluata, U. (1994). Circulation and hydrography of the Sea of Marmara. Progress in Oceanography, 34, 285–334. Bethoux, J. P. (1979). Budgets of the Mediterranean Sea. Their dependence on the local climate and on the characteristics of the Atlantic waters. Oceanologica Acta, 2, 157–163. Bethoux, J. P., Gentili, B., & Tailliez, D. (1998). Warming and freshwater budget change in the Mediterranean since the 1940s, their possible relation to the greenhouse effect. Geophys. Res. Let., 25, 1023–1026. Bethoux, J. P., & Gentili, B. (1999). Functioning of the Mediterranean Sea: past and present changes related to fresh water input and climatic changes. J. Mar. Sys., 20, 33–47. Bethoux, J. P., Gentili, B., Morin, P., Nicolas, E., Pierre, C., & 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. Progr. Oceanogr., 44(1–3), 131–146. Bethoux, J. P. (1980). Mean water fluxes across sections in the Mediterranean Sea, evaluated on the basis of water and salt budgets and of observed salinities. Oceanologica Acta, 3, 79–88. Bethoux, J. P., Gentili, B., Raunet, J., & Taillez, D. (1990). Warming trend in the Western mediterranean deep water. Nature, 347, 660–662. Bethoux, J. P., & Gentili, B. (1996). The Mediterranean Sea, coastal and deep-sea signatures of climatic and environmental changes. Journal of Marine Systems, 7, 383–394. Bignami, F., Marullo, S., Santoleri, R., & Schiano, M. E. (1995). Longwave radiation budget in the Mediterranean Sea. J. Geophys. Res., 100(C2), 2501–2504. Boscolo, R., & Bryden, H. L. (2001). Causes of long-term changes in Aegean Sea deep water. Oceanologica Acta, 24, 519–527. Boukthir, M., & Barnier, B. (2000). Seasonal and inter-annual variations in the surface freshwater flux in the Mediterranean Sea from the ECMWF re-analysis project. J. Mar. Sys., 24, 343–354. Brandt, P., Rubino, A., Sein, D. V., Baschek, B., Izquierdo, A., & Backhaus, J. O. (2004). Sea level variations in the Western Mediterranean studied by a numerical tidal model of the Strait of Gibraltar. J. of Phys. Oceanogr., 34(4), 433–443. Brankart, J. M., & Pinardi, N. (2001). Abrupt cooling of the Mediterranean Levantine intermediate water at the beginning of the 1980s: observational evidence and model simulation. J. Phys. Oceanogr., 31(8), 2307–2320. Bruce, J. G., & Charnock, H. (1965). Studies in winter sinking of cold water in the Aegean Sea. Rapport Comm. Int. Mer Medit., 18, 773–778. Brasseur, P. (1995). Free data offered to researchers studying the Mediterranean. EOS Transactions, American Geophysical Union, 76, 37, 363.
272 Mediterranean Climate Variability
Brasseur, P., Beckers, J. M., Brankart, J. M., & Schoenauen, R. (1996). Seasonal temperature and salinity fields in the Mediterranean Sea: climatological analyses of a historical data set. Deep Sea Res., 43, 159–192. Bryden, H. L., & Kinder, T. H. (1991). Steady two-layer exchange through the Strait of Gibraltar. Deep-Sea Research, 38(Suppl 1), S445–S463. Bryden, H. L., Candela, J., & Kinder, T. H. (1994). Exchange through the Strait of Gibraltar. Progress in Oceanography, 33, 201–248. Buljan, M., & Zore-Armanda, M. (1976). Oceanographical properties of the Adriatic Sea. Oceanogr. Mar. Biol., 14, 11–98. Bunker, A. F., Charnock, H., & Goldsmith, R. A. (1982). A note on the heat balance of the Mediterranean and Red Seas. Journal of Marine Research, 40, 73–84. Buongiorno Nardelli, B., & Salusti, E. (2000). On dense water formation criteria and their application to the Mediterranean Sea. Deep-Sea Research I, 47(Suppl 1), 193–221. Candela, J., Winant, C. D., & Bryden, H. L. (1989). Meteorologically forced subinertial flows through the Strait of Gibraltar. J. Geophys. Res., 94(C9), 12667–12679. Cazenave, A., Cabanes, C., Dominh, K., & Mangiarotti, S. (2001). Recent sea level changes in the Mediterranean Sea revealed by TOPEX/POSEIDON satellite altimetry. Geophys. Res. Let., 28(8), 1607–1610. Charnock, H. (1989). Temperature and salinity changes in the deep water of the Western Mediterranean Basin. 3rd POEM Scientific Workshop, Harvard, Cambridge, USA. Church J. A., Gregory, J. M., Huybrechts, P., Kuhn, M., Lambeck, K., Nhuan, M. T., Qin, D., & Woodworth, P. L. (2001). Chapter 11 Changes in sea level. Intergovernmental Panel on Climate Change Third Assessment Report, Cambridge University Press. CIESM (2002). Tracking long-term hydrological change in the Mediterranean Sea, Report of CIESM Workshop, Monaco, April 22–24, pp 9. Cullen, H. M., Kaplan, A., Arkin, P. A., & de Menocal, P. B. (2002). Impact of the North Atlantic Oscillation on Middle Eastern climate and stream flow. Clim. Change, 55, 315–338. De Putter, T., Loutre, M. F., & Wansard, G. (1998). Decadal periodicities of Nile River historical discharge (A. D. 622–1470) and climatic implications. Geoph. Res. Lett., 25(16), 3193–3196. Della Vedova, B., Foucher, J. P., Pellis, G., Harmegnies, F., & Medriff Consortium (1995). Heat flow measurements on the Mediterranean ridge indicate transient processes of heat transfer between the sediment and the water column. Rapport Commission internationale Mer Me´dite´rrane´e, 34, 100. Della Vedova, B., Pellis, G., Camerlenghi, A., Foucher, J. P., & Harmegnies, F. (2003). Thermal history of deep-sea sediments as a record of recent changes in the deep circulation of the Eastern Mediterranean. Journal of Geophysical Research, 108(C9), 8110, doi:10.1029/2002JC001402. Demirov, E., & Pinardi, N. (2002). Simulation of the Mediterranean Sea circulation from 1979 to 1993: Part I. The interannual variability. J. Marine Systems, 33–34, 23–50. de Madron, X. D., Zervakis, V., Theocharis, A., & Georgopoulos, D. (2005). Comments on Cascades of dense water around the world ocean, Progress in Oceanography, 64(1), 83–90.
Changes in the Oceanography of the Mediterranean Sea
273
Dunkeloh, A., & Jacobeit, J. (2003). Circulation dynamics of Mediterranean precipitation variability 1948–98. Int. J. Climatology, 23(15), 1843–1866. Elfatih, A., & Eltahir, B. (1996). El Nin˜o and the natural variability in the flow of the Nile River. Water Resour. Res., 32, 131–137. El-Gindy, A. A. H., El-Din, S., & H. Sh. (1986). Water masses and circulation patterns in the deep layer of the Eastern Mediterranean. Oceanologica Acta, 9, 239–248. Fekete, B. M., Vorosmarty, C. J., & Grabs, W. (1999). Global composite runoff fields on observed river discharge and simulated water balances, Tech. Rep., 22, Global Runoff Data Centre, Koblenz, Germany. Fenoglio-Marc, L. (2002). Long-term sea level change in the Mediterranean Sea from multi-sattelite altimetry and tide gauges. Physics and Chemistry of the Earth, 27, 1419–1431. Franco, P., Jeftic, L., Malanotte-Rizzoli, P., Michelato, A., & Orlic, M. (1982). Descriptive model of the Northern Adriatic. Oceanologica Acta, 5, 379–389. Fuda, J. L., Millot, C., Taupier-Letage, I., Send, U., & Bocognano, J. M. (2000). XBT monitoring of a meridian section across the Western Mediterranean Sea. Deep-Sea Res., 47, 2191–2218. Fuda, J.-L., Etiope, G., Millot, C., Favali, P., Calcara, M., Smriglio, G., & Boschi, E. (2002). Warming, salting and origin of the Tyrrhenian deep water. Geophysical Res. Lett., 29(19), 1898, doi: 10.1029/2001GL014072. Garrett, C., Outbridge, R., & Thompson, K. (1993). Interannual variability in Mediterranean heat and buoyancy fluxes. J. of Climate, 6, 900–910. Gasparini, G. P., Ortona, A., Budillon, G., Astraldi, M., & Sansone, E. (2005). The effect of the Eastern Mediterranean transient on the hydpographic characteristics in the Strait of Sicily and in the Tyrrhenian sea. Deep-Sea Res. I, 52, 915–935. Georgopoulos, D., Theocharis, A., & Zodiatis, G. (1989). Intermediate water formation in the Cretan Sea (S. Aegean Sea). Oceanologica Acta, 12(4), 353–359. Gertman, I., Pinardi, N., Popov, Y., & Hecht, A. (2005). Aegean sea water masses during the early stages of the Eastern Mediterranean climatic transient (1988–1990). J. Phys. Oceanogr., submitted. Gilman, C., & 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, 5119–5134. Grbec, B., Morovic, M., & Zore-Armanda, M. (2003). Mediterranean oscillation index and its relationship with salinity fluctuations in the Adriatic Sea. Acta Adriatica, 44, 61–76. Gregg, M. C., & O¨zsoy, E. (1999). Mixing on the Black Sea shelf north of the Bosphorus. Geophysical Research Letters, 26, 1869–1872. Gregg, M. C., O¨zsoy, E., & Latif, M. A. (1999). Quasi-Steady exchange flow in the bosphorus. Geophys. Res. Lett., 26, 83–86. Gunnerson, C. G., & Ozturgut, E. (1974). The Bosphorus. In: E. T. Degens, & D. A. Ross (Eds), The Black Sea – Geology, Chemistry and Biology (American Association of Petroleum Geologists, Tulsa, p. 103). Hagemann, S., Arpe, K., & Bengtsson, L. (2005). Validation of the hydrological cycle of ERA40, ECMWF ERA-40 Proj. Rep. Ser. 24, Reading, UK. Hamza, W., Ennet, P., Tamsalu, R., & Zalensky, V. (2003). The 3D physical biological model study in the Egyptian Mediterranean coastal area. Aquatic Ecology, 37, 307–324.
274 Mediterranean Climate Variability
Hecht, A., & Gertman, I. (2001). Physical features of the Eastern Mediterranean resulting from the integration of POEM data with Russian Mediterranean cruises. Deep-Sea Research I, 48, 1847–1876. Herbaut, C., Martel, F., & Crepon, M. (1997). A sensitivity study of the general circulation of the Western Mediterranean Sea. 2. The response to atmospheric forcing. J. Phys. Oceanogr., 27(10), 2126–2145. Hopkins, T. S. (1978). Physical processes in the mediterranean basins. In: B. Kjerfve (Ed), Estuarine Transport Processes, (University of South Carolina Press, Columbia, SC, pp. 285–286). Hopkins, T. S. (1985). Physics of the sea. In: R. Margalef (Ed), Key Environments – Western Mediterranean, (Pergamon Press, Oxford, pp. 116–117). Hopkins, T. S. (1999). The thermohaline forcing of the Gibraltar exchange. Journal of Marine Systems, 20, 1–31. Hurrell, J. W. (1995). Decadal trends in the North Atlantic Oscillation regional temperatures and precipitation. Science, 269, 676–679. Ibanez, C., Prat, N., & Canicio, A. (1996). Changes in the hydrology and sediment transport produced by large dams on the lower Ebro river and its estuary. Reg. Riv. Res. & Man., 12, 51–62. Jacobeit, J. (2005). The Eastern Mediterranean transient in relation to atmospheric circulation dynamics, EGU General Assembly, Vienna, April 2005, EGU05-A-07108. Josey, S. A. (2003). Changes in the heat and freshwater forcing of the Eastern Mediterranean and their influence on deep water formation. Journal of Geophysical Research, 108(C7), 3237, doi:10.1029/2003JC001778. Josey, S. A., Kent, E. C., & Taylor, P. K. (1999). New insights into the ocean heat budget closure problem from analysis of the SOC air-sea flux climatology. J. Climate, 12, 2880. Kahya, E., & Karabo¨rk, M. C¸. (2001). The analysis of El Nin˜o and La Nin˜a signals in streamflows of Turkey. Int. J. Climatol., 21, 1231–1250. Kara, A. B., Wallcraft, A. J., & Hurlburt, H. (2005). Sea surface temperature sensitivity to water turbidity from simulations of the turbid black sea using HYCOM. J. Phys. Oceanogr., 35, 33–54. Klein, B, Roether, W., Manca, B. B., Bregant, D., Beitzel, V., Kovacevic, V., & Luchetta, A. (1999). The large deep water transient in the Eastern Mediterranean. Deep-Sea Research I, 46, 371–414. Klein, B., Roether, W., Civitarese, G., Gacˇic´, M., Manca, B. B., & Ribera D’alcala, M. (2000). Is the Adriatic returning to dominate the production of Eastern Mediterranean deep water? Geophysical Research Letters, 27(20), 3377–3380. Konovalov, S. K., Luther, G. W. III, Friederich, G. E., Nuzzio, D. B., Tebo, B. M., Murray, J. W., Oguz, T., Glazer, B., Trouwborst, R. E., Clement, B., Murray, K. J., & Romanov, A. S. (2003). Lateral injection of oxygen with the Bosporus plume: fingers of oxidizing potential in the Black Sea. Limnology and Oceanography, 48(6), 2369–2376. Kovacevicc´, V., Gacic, M., & Poulain, P.-M. (1999). Eulerian current measurements in the Strait of Otranto and in the Southern Adriatic. Journal of Marine Systems, 20, 255–278. Krahmann, G., & Schott, F. (1998). Longterm increases in Western Mediterranean salinities and temperatures: anthropogenic and climatic sources. Geophysical Research Letters, 25, 4209–4212.
Changes in the Oceanography of the Mediterranean Sea
275
Lacombe, H., & Tchernia, P. (1958). Tempe´rature et salinite´ profondes en Me´diterane´e en pe´riode d’ e´te´ (fin juin–octobre). Bulletin information comite´ central d’oce´anographie et d’ e´tude des coˆtes, 10. Lacombe, H., & Tchernia, P. (1960). Quelques traits ge´ne´raux de l’hydrologie Me´diterrane´ene. Cahiers Oce´anographiques, 20(8), 528–547. Lacombe, H., Tchernia, P., & Gamberoni, L. (1985). Variable bottom water in the Western Mediterranean basin. Progress in Oceanography, 14, 319–338. La Fuente, J. G., Delgado, J., Vargas, J. M., Vargas, M., Plaza, F., & Sarhan, T. (2002). Low-frequency variability of the exchanged flows through the Strait of Gibraltar during CANIGO. Deep-Sea Research II, 49, 4051–4067. Larnicol, G., LeTraon, P. Y., Ayoub, N., & DeMey, P. (1995). Mean sea level and surface circulation variability of the Mediterranean Sea from 2 years of TOPEX/ POSEIDON altimetry. J. Geophys. Res., 100(C12), 25163–25177. Lascaratos, A. (1993). Estimation of deep and intermediate water mass formation rates in the Mediterranean Sea. Deep-Sea Research II, 40, 1327–1332. Lascaratos, A., & Nittis, K. (1998). A high resolution 3-D numerical study of intermediate water formation in the Levantine Sea. Journal of Geophysical Research, 103, 18497–18511. Lascaratos, A., Williams, R., & Tragou, E. (1993). A mixed layer study of the formation of Levantine intermediate water. Journal of Geophysical Research, 98, 739–749. Lascaratos, A., Roether, W., Nittis, K., & Klein, B. (1999). Recent changes in deep water formation and spreading in the Mediterranean Sea: a review. Progress in Oceanography, 44(1–2), 36. Latif, M. A., Ozsoy, E., Oguz, T., & Unluata, U. (1991). Observations of the Mediterranean inflow into the Black Sea. Deep-Sea Res., 38(Suppl. 2), 711–723. Leaman, K. D., & Schott, F. A. (1991). Hydrographic structure of the convection regime in the Gulf of Lions: winter 1987. Journal of Physical Oceanography, 21, 575–598. Lee, B.-S., Bullister, J. L., Murray, J. W., & Sonnerup, R. E. (2002). Anthropogenic chlorofluorocarbons in the Black Sea and the Sea of Marmara. Deep-Sea Research, Part I, 49(5), 895–913. Lionello, P. (2005). Extreme surges in the Gulf of Venice. Present and future climate. In: C. Fletcher, & T. Spencer (Eds), Venice and its lagoon, State of knowledge, (Cambridge University Press, Cambridge, UK, pp. 59–65). Lionello, P., & Sanna, A. (2005). Mediterranean wave climate variability and its links with NAO and Indian monsoon, Clim. Dyn., 25(6), 611–623. LIWEX Group (2003). The Levantine intermediate water experiment (LIWEX) group: Levantine basin – a laboratory for multiple water mass formation processes. J. Geophys. Res., 198, doi:10.1029/2002JC001643. Macdonald, A., Candela, J., & Bryden, H. L. (1994). An estimate of the net heat transport through the Strait of Gibraltar. In: P. E. La Violette (Ed), Coastal Estuarine Studies (Washington, D. C., pp. 13–32). Maillard, C., Balopoulos, E., Fichaut, M., Giorgetti, A., Iona, A., Latrouite, A., & Manca, B. (2001). Mater data manual version 3, Vol. 1: Data management structure, inventories, formats and codes, quality assurance. Rap. Int. IFREMER/TMSI/IDM/ SISMER/01-058, 67 pp. Malanotte-Rizzoli, P. (1991). The Northern Adriatic Sea as a prototype of convection and water mass formation on the continental shelf. In: P. C. Chu, & J. C. Gascard
276 Mediterranean Climate Variability
(Eds), Deep Convection and Deep Water Formation in the Oceans. Elsevier Oceanogr. Ser., 57, 229–239. Malanotte-Rizzoli, P., & Hecht, A. (1988). Large-scale properties of the eastern Mediterranean: a review. Oceanologica Acta, 11, 323–335. Malanotte-Rizzoli, P., Manca, B. B., Ribera Dacala, M., Theocharis, A., Bergamasco, A., Bregant, D., Budillon, G., Civitarese, G., Georgopoulos, D., Michelato, A., Sansone, E., Scarazzato, P., & Souverrmezoglou, E. (1997). A synthesis of the ionian hydrography, circulation and water mass pathways during POEM— Phase I. Progress in Oceanography, 39, 153–204. Malanotte-Rizzoli, P., Manca, B. B., d’alcala, M. R., Theocharis, A., Brenner, S, Budillon, G., & O¨zsoy, E. (1999). The Eastern Mediterranean in the 80s and in the 90s: the big transition in the intermediate and deep circulations. Dynamics of Atmospheres and Oceans, 29, 365–395. Manca, B., & Giorgetti, A. (1998). Thermohaline properties and circulation patterns in the Southern Adriatic Sea from May (1995) to February 1996. In: M. Piccazzo (Ed.), Atti 12 Congr. Associazione Italiana di Oceanologia e Limnologia, Isola di Vulcano, 1996, vol. II, A.I.O.L., Genova, pp. 399–414. Manca, B. B., Kovacevic, V., Gacic, M., & Viezzoli, D. (2002). Dense water formation in the Southern Adriatic Sea and spreading into the Ionian Sea in the period 1997–1999. Journal of Marine Systems, 33–34, 133–154. Manca, B. B., Budillon, G., Scarazzato, P., & Ursella, L. (2003). Evolution of dynamics in the Eastern Mediterranean affecting water mass structures and properties in the Ionian and Adriatic Seas. J. Geophys. Res., 108(C9), art. no. 8102. Manca, B., Burca, M., Giorgetti, A., Coatanoan, C., Garcia, M.-J., & Iona, A. (2004). Physical and biochemical averaged vertical profiles in the Mediterranean regions: an important tool to trace the climatology of water masses and to validate incoming data from operational oceanography. Journal of Marine Systems, 48(1–4), 83–116. Margat J. (1992). L’eau dans le bassin me´diterrane´en. Vol. 6, Economica, 196 pp. Mariotti, A., Struglia, M. V., Zeng, N., & Lau, K.-M. (2002). The hydrological cycle in the Mediterranean region and implications for the water budget of the Mediterranean. Sea. J. Climate, 15, 1674–1690. Marsigli, L. F. (1681). Osservazioni lntomo al Bosforo Tracio o vero Canale di Constantinopli rappresentate in lettera alla Sacra Real Maesta di Cristina Regina di Svezia Nicolb Angelo Tinassi, Rome. Reprinted in: Bollettino di pesca, di piscicoltura e di idrobiologia. Anno Xl, Fast. 5, Settembre–Ottobre 1935, 734–758. Martin, J. M., & Milliman, J. D. (1997). EROS 2000 European River Ocean System. The Western Mediterranean: an introduction. Deep-Sea Research, 44, 521–529. MEDATLAS Group (1997). MEDATLAS 1997 database and climatological atlas of temperature and salinity. Three CD-ROMs and WWW server, IFREMER Ed. Mertens, C., & Schott, F. (1998). Interannual variability of deep water formation in the northwestern Mediterranean Sea. J. Phys. Oceanogr., 28, 1410–1424. Miller, A. R. (1974). Deep convection in the Aegean Sea, in Processus de Formation des Eaux Oce´aniques Profondes. Colloques Internationaux du C. N. R. S. 215. Miller, L., & Douglas, B. C. (2004). Mass and volume contributions to twentieth-century global sea level rise. Nature, 428, 406–409. Millot, C. (1987). Circulation in the Western Mediterranean Sea. Oceanologica Acta, 10(2), 143–149.
Changes in the Oceanography of the Mediterranean Sea
277
Millot, C. (1999). Circulation in the Western Mediterranean Sea. J. Marine Systems, 20(1–4), 423–444. Millot, C., Benzohra, M., & Taupier Letage, I. (1997). Circulation off Algeria inferred from the Mediprod-5 current meters. Deep Sea Res. I, 44(9–10), 1467–1481. Millot, C., & Taupier-Legate, I. (2005). Additional evidence of LIW entrainment across the Algerian subbasin by mesoscale eddies and not by a permanent westward flow. Progress in Oceanography, 66(2–4), 231–250. doi: 10.1016. Morcos, S. A. (1972). Sources of Mediterranean intermediate water in the Levantine Sea. In: A. L. Gordon (Ed), Studies in Physical Oceanography, Vol. 2 (Gordon and Breach, New York, pp. 185–206). Nielsen, J. N. (1912). Hydrography of the Mediterranean and adjacent seas. Danish Oceanographic Expeditions 1908–10. Report I, pp. 72–191. Nittis, K., & Lascaratos, A. (1988). Diagnostic and prognostic numerical studies of LIW formation. Journal of Marine Systems, 18, 179–195. Nittis, K., Lascaratos, A., & Theocharis, A. (2003). Dense water formation in the Aegean Sea: numerical simulations during the Eastern Mediterranean Transient. J. Geophys. Res., 108(C9), 8120. Ogu¨z, T., & Malanotte-Rizzoli, P. (1996). Seasonal variability of wind and thermohaline driven circulation in the Black Sea: modeling studies. J. Geophys. Res., 101, 16551–16569. Ogu¨z, T., Cokacar, T., Malanotte-Rizzoli, P., & Ducklow, H. W. (2003). Climatic warming and accompanying changes in the ecological regime of the Black Sea during 1990s. Global Biogeochemical Cycles, 17(3), 1088, doi:10.1029/2003GB002031. Orlic, M., & Pasaric, M. (2000). Sea-level changes and crustal movements recorded along the east Adriatic coast. II Nuovo Cimento, 23, 351–364. Ovchinnikov, I. M. (1974). On the water balance of the Mediterranean Sea. Oceanology, 14, 198–202. Ovchinnikov, I. M. (1984). Formation of intermediate waters in the Mediterranean Sea. Oceanology, 24, 217–255. Ovchinnikov, I. M., & Plakhin, Ye. A. (1984). Formation of the intermediate waters of the Mediterranean Sea in the Rhodes Cyclonic Gyre. Oceanology, 24, 317–319. Ovchinnikov, I. M., Zets, V. I., Kroshaye, V. G., Menirovsk, M. S., & Udodov, A. I. (1987). Winter convection in the Adriatic and formation of deep Eastern Mediterranean water. Oceanoglogica Acta, 5B, 89–92. Ovchinnikov, I. M., Popov, Yu. I., & Gertman, I. F. (1990). Investigation of the formation of deep waters in the Eastern Mediterranean Sea during the 36th cruise of the R/V Ya. Gakkel’ 22 January–13 April (1990). Oceanology, 30(6), 769–771. O¨zturgut, E. (1976). The source and spreading of the Levantine intermediate water in the Eastern Mediterranean. Saclant antisubmarine warfare Res. Cen. Memor. SM-92, La Spezia, Italy, pp. 45. Painter, S. C., & Tsimplis, M. N. (2003). Temperature and salinity trends in the upper waters of the Mediterranean Sea as determined from the MEDATLAS dataset. Continental Shelf Research, 23, 1507–1522. Pasaric, M., & Orlic, M. (2001). Long-term meteorological preconditioning of the North Adriatic coastal floods. Cont. Shelf Research, 21(3), 263–278. Paskoff, R. P. (2004). Potential implications of sea level rise for France. J. Coastal Res., 20(2), 424–434.
278 Mediterranean Climate Variability
Perkins, H., Kinder, T., & La Violette, P. (1990). The Atlantic inflow in The Western Alboran Sea. Journal of Physical Oceanography, 20(2), 242–263. Pinardi, N., Korres, G., Lascaratos, A., Roussenov, V., & Stanev, E. (1997). Numerical simulation of the interannual variability of the Mediterranean Sea upper ocean circulation. Geophysical Research Letters, 24(4), 425–428. Pinardi, N., Arneri, E., Crise, A., Ravaioli, M., & Zavatarelli, M. (2005). The physical, sedimentary and ecological structure and variability of shelf areas in the Mediterranean, The Sea, 14, in press. Pirazzoli, P. A., & Tomasin, A. (2002). Recent evolution of surge-related events in the Northern-Adriatic Sea. J. Coast. Res., 18(3), 537–554. Pirazzoli, P. A., & Tomasin (2003). Recent near-surface wind changes in the Central Mediterranean and Adriatic areas. Int. J. Climatology, 23(8), 963–973. POEM GROUP (1992). General circulation of the Eastern Mediterranean. Earth-Science Review, 32, 285–309. Pollak, M. I. (1951). The sources of deep water in the Eastern Mediterranean sea. Journal of Marine Research, 10, 128–152. Poulain, P.-M. (2001). Adriatic Sea surface circulation as derived from drifter data between 1990 and 1999. J. Marine Sys., 29, 3–32. Raicich, F. (1996). On the fresh water balance of the Adriatic Sea. J. Mar. Sys., 9, 305–319. Raicich, F. (2003). Recent evolution of sea-level extremes at Trieste (Northern Adriatic). Continental Shelf Research, 23, 225–235. Reihn, M. (1995). Deep water formation in the Western Mediterranean. J. Geophys. Res., 100, 6943–6959. Rixen, M., Beckers, J.-M., Levitus, S., Antonov, J., Boyer, T., Maillard, C., Fichaut, M., Balopoulos, E., Iona, S., Dooley, H., Garcia, M. J., Manca, B., Giorgetti, A., Manzella, G., Mikhailov, N., Pinardi, N., Zavatarelli, M., & the Medar consortium (2005). The Western Mediterranean deep water: a proxy for climate change. Geophys. Res. Lett., 32(12), Art. No. L12608. Robinson, A. R., Golnaraghi, M., Leslie, N., Artegiani, A., Hecht, A., Lazzone, E., Michelato, A., Sanzone, E., Theocharis, A., & Unluata, U. (1991). Structure and variability of the Eastern Mediterranean general circulation. Dynamics of Atmospheres and Oceans, 15, 215–240. Roether, W., & Klein, B. (1998). The great Eastern Mediterranean deep-water transient. Rapport Commission internationale Mer Me´dite´rrane´e, 35, 12–16. Roether, W., & Schlitzer, R. (1991). Eastern Mediterranean deep water renewal on the basis of chlorofluoromethane and tritium data. Dynamics of Atmospheres and Oceans, 15(3–5), 333–354. Roether, W., Roussenov, V. M., & Well, R. (1994). A tracer study of the thermohaline circulation of the Eastern Mediterranean. In: P. Malanotte-Rizzoli, & A. R. Robinson (Eds.), Ocean Processes in Climate Dynamics: Global and Mediterranean Examples, Vol. 419, NATO ASI Series C (Kluwer Academic Publisher, pp. 371–394). Roether, W., Manca, B. B., Klein, B., Bregant, D., Georgopoulos, D., Beitzel, V., Kovacevic, V., & Luchetta, A. (1996). Recent changes in Eastern Mediterranean deep waters. Science, 271, 333–335. Rohling, E. J., & Bryden, H. L. (1992). Man-induced salinity and temperature increases in Western Mediterranean deep water. Journal of Geophysical Research, 97, 11191–11198.
Changes in the Oceanography of the Mediterranean Sea
279
Ross., T., Garrett, C., & Le Traon, P.-Y. (2000). Western Mediterranean sea-level rise: changing exchange flow through the Strait of Gibraltar. Geophys. Res. Let., 27, 2949–2952. Roussenov, V., Stanev, E., Artale, V., & Pinardi, N. (1995). A seasonal model of the Mediterranean Sea general circulation. Journal of Geophysical Research, 100, 13515–13538. Rupolo, V., Marullo, S., & Iudicone, D. (2003). Eastern Mediterranean transient studied with Lagrangian diagnostics applied to a Mediterranean OGCM forced by satellite SST and ECMWF wind stress for the years 1988–1993. J. Geophys. Res., 108(C9), 8121, doi:10.1029/2002JC001403. Samuel, S. L., Haines, K., Josey, S. A., & Myers, P. G. (1999). Response of the Mediterranean Sea thermohaline circulation to observed changes in the winter wind stress field in the period 1980–93. J. Geophys. Res., 104, 5191–5210. Schiano, M. E. (1996). Insolation over the Western Mediterranean Sea: a comparison of direct measurements and Reed’s formula. J. Geophys. Res., 101, 3831–3838. Schlitzer, R., Roether, W., Oster, H., Junghans, H., Hausmann, M., Johannsen, H., & Michelato, A. (1991). Chlorofluoromethane and oxygen in the Eastern Mediterranean. Deep-Sea Research, 38, 1531–1535. Schott, G. (1915). Die Gewasser des Mittelmeeres. Annal. Der Hydrographie und maritimen Meteorologie. Deutsche Seewarte, 43, 1–18, 63–79. Send, U., Schott, F., Gaillard, F., & Desaubies, Y. (1995). Observation of deep water convection regime with acoustic tomography. J. Geophys. Res., 100, 6927–6941. Send, U., Font, J., Krahmann, G., Millot, C., Rhein, M., & Tintore, J. (1999). Recent advances in observing the physical oceanography of the Western Mediterranean Sea. Progress in Oceanogr., 44, 37–64. Shorthouse, C. A., & Arnell, N. W. (1997). Spatial and temporal variability in European river flows and the North Atlantic Oscillation, Proc. FRIEND ’97-Regional hydrology: concepts and models for sustainable water resource management Postojna, Slovenia, IAHS Publ. No. 246, 77–85. Simmons, A. J., & Gibson, J. K. (2000). The ERA-40 project plan. ERA-40 Project Report Series, Tech. Rep. 1, European centre for medium-range weather forecasts, Reading, UK. A. I. Simonov, & E. N. Altman (Eds) (1991). Hydrometeorology and hydrochemistry of the USSR Seas. The Black Sea, Vol. IV Gidrometeoizdat, St. Petersburg, pp. 430. Skliris, N., & Lascaratos, A. (2004). Impacts of the Nile river damming on the thermohaline circulation and water mass characteristics of the Mediterranean Sea. Journal of Marine Systems, 52, 121–143. Smith, R. L. (1986). Extreme value theory based on the n-largest annual events. J. Hydrology, 86, 27–43. Stanev, E. V. (1990). On the mechanisms of the Black Sea circulation. Earth-Science Reviews, 28, 285–319. Stanev, E., & Peneva, E. L. (2002). Regional sea level response to global climatic change: Black Sea examples. Global and Planetary Changes, 32, 33–47. Stanev, E. V., Le Traon, P. Y., & Peneva, E. L. (2000). Sea level variations and their dependency on meteorological and hydrological forcing: analysis of altimeter and surface data for the Black Sea. J. Geoph. Res., 105(C7), 17203–17216. Stanev, E. V., Simeonov, J. A., & Peneva, E. L. (2001). Ventilation of Black Sea pycnocline by the Mediterranean plume. Journal of Marine Systems, 31, 77–97.
280 Mediterranean Climate Variability
Staneva, J. V., Dietrich, D., Stanev, E. V., & Bowman, M. (2001). Rim current and coastal eddy mechanisms in an eddy-resolving Black Sea general circulation model. J. Marine Sys., 3, 137–157. Stanev, E. V., Staneva, J., Bullister, J. L., & Murray, J. W. (2004). Ventilation of the Black Sea pycnocline. Parameterization of convection, numerical simulations and validations against observed chlorofluorocarbon data. Deep-Sea Research I, 51, 2137–2169. Stommel, H. (1972). Deep winter-time convection in the Western Mediterranean Sea. In: A. L. Gordon (Ed), Studies in Physical Oceanography, Vol. 2 (Gordon and Breach, New York, pp. 207–218). Stommel, H., Bryden, H., & Mangelsd, P. (1973). Does some of Mediterranean outflow come from great depth. Pure and Applied Geophysics, 105(4), 879–889. Stratford, K., & Haines, K. (2002). Modelling changes in Mediterranean thermohaline circulation 1987–1995. J. Marine Systems, 33–34, 51–62. Struglia, M. V., Mariotti, A., & Filograsso, A. (2004). River discharge into the Mediterranean Sea: climatology and aspects of the observed variability. J. Climate, 17, 4740–4751. The Liwex Group (2003). The Levantine Intermediate Water Experiment (Liwex) Group, Levantine basin—a laboratory for multiple water mass formation processes. Journal of Geophysical Research, 108(C9), 8101, doi:10.1029/2002JC001643. Theocharis, A., & Georgopoulos, D. (1993). Dense water formation over the Samothraki and Lemnos plateaux in the North Aegean Sea (Eastern Mediterranean Sea). Continental Shelf Research, 13(8–9), 919–939. Theocharis, A, Georgopoulos, D., Lascaratos, A., & Nittis, K. (1993). Water masses and circulation in the Central region of the Eastern Mediterranean: Eastern Ionian, South Aegean and Northwest Levantine, 1986–1987. Deep Sea Research II, 40, 1121–1142. Theocharis, A., Gacic, M., & Kontoyiannis, H. (1998). Physical and dynamical processes in the coastal and shelf areas of the Mediterranean. In: A. R. Robinson, & K. H. Brink (Eds), The Sea, Vol. 11 (John Wiley and Sons, Inc., pp. 863–887). Theocharis, A., Balopoulos, E., Kioroglou, S., Kontoyiannis, H., & Iona, A. (1999a). A synthesis of the circulation and hydrography of the South Aegean Sea and the Straits of the Cretan Arc (March 1994–January 1995). Progress in Oceanography, 44, 469–509. Theocharis, A., Nittis, K., Kontoyiannis, H., Papageorgiou, E., & Balopoulos, E. (1999b). Climatic changes in the Aegean Sea influence the Eastern Mediterranean thermohaline circulation. Geophys. Res. Lett., 26, 1617–1620. Theocharis, A., Klein, B., Nittis, K., & Roether, W. (2002). Evolution and status of the Eastern Mediterranean transient (1997–1999). J. Marine Systems, 33–34, 91–116. Tixeront, J. (1970). Le bilan hydrologique de la mer Noire et de la mer Me´diterrane´e. Cah. Oceanogr., 22(3), 227–237. Tolmazin, D. (1985). Changing coastal oceanography of the Black Sea. I: Northwestern shelf. Progress in Oceanography, 15, 217–276. Tragou, E., & Lascaratos, A. (2003). The role of aerosols on the Mediterranean solar radiation. J. Geophys. Res., 108(C2), doi:10.1029/2001JC001258. Tragou, E., Josey, S., Zervakis, V., Maderich, V., & Theocharis, A. (2005). Identifying signatures of dense water formation events over the Aegean Sea in the last fifty years, EGU General Assembly, Vienna, 2005, EGU05-A-08769.
Changes in the Oceanography of the Mediterranean Sea
281
Trigo, I. F., & Davies, T. D. (2002). Meteorological conditions associated with sea surges in Venice: a 40 year climatology. Int. J. Climatology, 22(7), 787–803. Trigo, R. M., Pozo-Vazquez, D., Osborn, T. J., Castro-Diez, J., Gamiz-Fortis, S., & Esteban-Parra, M. J. (2004). North Atlantic Oscillation influence on precipitation, river flow and water resources in the Iberian Peninsula. Int. J. Climatol., 24, 925–944. Tsimplis, M. N., & Blackman, D. (1997). Extreme sea level listribution and return periods in the Aegean and Ionian Seas. Estuarine Coastal and Shelf Science, 44, 79–89. Tsimplis, M. N., & Spencer, N. E. (1997). Collection and analysis of monthly mean sea level data in the Mediterranean and Black Sea. Journal of Coastal Research, 13(2), 534–544. Tsimplis, M. N., Velegrakis, A. F., Theocharis, A., & Collins, M. B. (1997). Low frequency current variability at the straits of Crete, Eastern Mediterranean. Journal of Geophysical Research, 102(C11), 25005–25020. Tsimplis, M. N., Velegrakis, A. F., Drakopoulos, P. G., Theocharis, A., & Collins, M. B. (1999). Cretan deep water outflow in the Eastern Mediterranean. Progress in Oceanography 44(4), 531–551. Tsimplis, M. N., & Bryden, H. L. (2000). Estimation of the transports through the Strait of Gibraltar. Deep-Sea Research I, 47, 2219–2242. Tsimplis, M. N., & Baker, T. F. (2000). Sea level drop in the Mediterranean Sea: an indicator of deep water salinity and temperature changes? Geophysical Research Letters, 27(12), 1731–1734. Tsimplis, M. N., & Josey, S. A. (2001). Forcing of the Mediterranean Sea by atmospheric oscillations over the North Atlantic. Geophys. Res. Let., 28(5), 803–806. Tsimplis, M. N., & Rixen, M. (2002). Sea level in the Mediterranean Sea: The contribution of temperature and salinity changes. Geophys. Res. Let., 29(3), art no. 2136. Tsimplis, M. N., Josey, S. A., Rixen, M., & Stanev, E. V. (2004). On the forcing of sea level in the Black Sea. J. Geophys. Res. (Oceans), 109(C8), art. no. C08015. Tsimplis, M. N., A´lvarez-Fanjul, E., Gomis, D., Fenoglio-Marc, L., & Pe´rez, B. (2005a). Mediterranean sea level trends: atmospheric pressure and wind contribution. Geophys. Res. Lett., 32(20), Art. No. L20602. Tsimplis, M. N., Gomis, D., & Fenoglio-Marc, L. (2005b). Hydraulic jumps, deep water formation and sea level variability in the Mediterranean Sea, in preparation. Turley, M. (1999). The changing Mediterranean Sea – a sensitive ecosystem? Progress in Oceanography, 44, 387–400. Tziperman, E., & Speer, K. (1994). A study of water mass transformation in the Mediterranean Sea – analysis of climatological data and a simple 3-box model. Dynamics of Atmospheres and Oceans, 21(2–3), 53–82. Vigo, I., Garcı´ a, D., & Chao, B. F. (2005). Change of sea level trend in Mediterranean and Black Seas. Journal of Marine Research, in press. Vilibic, I., Domijan, N., Orlic, M., Leder, N., & Pasaric, M. (2004). Resonant coupling of a traveling air pressure disturbance with the east Adriatic coastal waters. J. Geophys. Res. (Oceans), 109(C10), art. no. C10001. Wadie, W. F. (1984). The effect of regulation of the Nile River discharge on the oceanographic conditions and productivity of the southeastern part of the Mediterranean Sea. Acta Adriatica, 25, 29–43. Waitz, J. S. (1755). v. Undersokning om orsaken, hvarfo¨re vattnet i Atlantiska hafvet altid stro¨mar in uti medelhafvet genom Sundet vid Gibraltar. Kungliga Svensk.a Vetenskupsakademiens Handlingar, 16, 27–50.
282 Mediterranean Climate Variability
Woodworth, P. L. (2003). Some comments on the long sea level records from the Northern Mediterranean. Journal of Coastal Research, 19, N. 1. Woodworth, P. L., & Blackman, D. L. (2004). Evidence for systematic changes in extreme high waters since the mid-1970s. J. Climate, 17(6), 1190–1197. Woolf, D., Shaw, A., & Tsimplis, M. N. (2003). The influence of the North Atlantic Oscillation on sea level variability in the North Atlantic region. The Global Atmosphere and Ocean System, 9(4), 145–167. Wu, P., & Haines, K. (1996). Modeling the dispersal of Levantine Intermediate Water and its role in Mediterranean deep water formation. J. Geophys. Res., 101, 6591–6607. Wu, P., Haines, K., & Pinardi, N. (2000). Toward an understanding of deep-water renewal in the Eastern Mediterranean. Journal of Physical Oceanography, 30, 443–458. Wu¨st, G. (1961). On the vertical circulation of the Mediterranean Sea. Journal of Geophysical Research, 66, 3261–3271. Yuce, H. (1996). Mediterranean water in the Strait of Istanbul (Bosphorus) and the Black Sea exit. Estuarine, Coastal and Shelf Science, 43(5), 597–616(20). Zavatarelli, M. (1999). Mediterranean Sea multiscale variability and environmental management issues: a scientific perspective. Progress in Oceanography, 44, 401–409. Zavatarelli, M., Raicich, F., Bregant, D., Russo, A., & Artegiani, A. (1998). Climatological biogeochemical characteristics of the Adriatic Sea. J. Mar. Sys., 18, 227–263. Zervakis, V., Georgopoulos, D., & Drakopoulos, P. G. (2000). The role of the North Aegean in triggering the recent Eastern Mediterranean climatic changes. J. Geophys. Res., 105, 26103–26116. Zervakis, V., & Georgopoulos, D. (2002). Hydrology and circulation in the North Aegean (Eastern Mediterranean) throughout 1997 and 1998. Mediterranean Marine Science, 3, 5–19. Zervakis, V., Georgopoulos, D., Karageorgis, A. P., & Theocharis, A. (2004). On the response of the Aegean Sea to climatic variability: a review. Int. J. Climatology, 24, 1845–1858, doi: 10.1002/joc.1108. Zodiatis, G. (1992). On the seasonal variability of the water masses circulation in the NW Levantine Basin-Cretan Sea and flows through the Eastern Cretan Straits. Annales Geophysicae, 10, 12–24.
Chapter 5
The Atlantic and Mediterranean Sea as Connected Systems Vincenzo Artale,1 Sandro Calmanti,1 Paola Malanotte-Rizzoli,2 Giovanna Pisacane,1 Volfango Rupolo1 and Mikis Tsimplis3 1
ENEA (Ente Nazionale per le Nuove Tecnologie l’Energia l’Ambiente) Roma, Italy (
[email protected],
[email protected],
[email protected],
[email protected]) 2 MIT (Massachusetts Institute of Technology), Boston, USA (
[email protected]) 3 SOC (Southampton Oceanography Centre), University of Southampton, UK (
[email protected])
5.1. Introduction The Mediterranean Sea is a mid-latitude marginal sea with a maximum depth of over 4,000 m and a limited exchange with the global ocean. Within the present climate, the Mediterranean Sea produces dense, warm and salty water that outflows through the Gibraltar Strait into the North Atlantic. The outflow amounts to about 1 Sv of water that can be over 5 C warmer than any other water mass in the North Atlantic at the same latitude and depth, and more than 1 psu saltier. After mixing with the surrounding water masses, the Mediterranean Outflow Water (MOW) is neutrally buoyant at about 1,000 m depth. The spreading of the salinity anomaly associated with the presence of MOW has captured much attention in the past, and yet, major uncertainties remain. A contribution to the average salinity of the world ocean equivalent to that of MOW would be achieved by distributing over the North Atlantic the net evaporation observed over the Mediterranean Sea. The present estimate of the
284 Mediterranean Climate Variability
freshwater budget of the North Atlantic (Ganachaud and Wunsch, 2003 and references therein) is rather uncertain, ranging from 0.2 to 0.8 Sv of net loss, north of 30 S. The corresponding estimate of the Mediterranean Water deficit ranges from 378 to 950 mm y 1 and therefore this also has a large error bar (Mariotti et al., 2002). Assuming an area for the Mediterranean Sea of 2.5 1012 m2 these values correspond to 0.03–0.08 Sv of net evaporation. Therefore the contribution of the Mediterranean Outflow to the freshwater budget of the North Atlantic can be currently estimated between 4 and 40%. Such a large uncertainty is quite impressive. It means that we are missing a fundamental quantitative detail concerning one of the important features of the climate system, that is the freshwater budget of the North Atlantic (Rahmstorf, 1996). In the broad sense discussed above, the North Atlantic plus the Mediterranean Sea can be viewed as a unique system whose ‘‘internal’’ dynamics, regulated by the exchanges at the Strait of Gibraltar, is still rather unknown. In this chapter, we mostly discuss numerical modelling results, as a complement to the data analysis presented in the previous chapter. We will discuss three issues related to the dynamics of such a system. We start focusing on the variability of water mass transformation processes inside the Mediterranean Sea with special attention to the time necessary for water masses formed within the Mediterranean Sea to spread into the North Atlantic. We then discuss the physics of the interface between the two sub-systems, which is the exchange at the Strait of Gibraltar. In the last section, using results from a hierarchy of numerical ocean models, we review the spreading of MOW in the North Atlantic and its possible relation to the variability of the meridional overturning circulation of the global ocean.
5.2. Mediterranean Outflow vs. Mediterranean Internal Variability In this section, we consider the water mass transformation processes occurring inside the Mediterranean Sea. In particular, we review the main pathways that compose the lower branch (the intermediate and deep circulation) of the Mediterranean Thermohaline Cell (THC) with the aim of identifying the regions where the properties of the pure Levantine Intermediate Water (LIW) are modified by mixing processes with other locally formed water masses. Then, we will discuss the characteristic time of propagation of passive tracers from the Eastern Mediterranean sub-basin (EM) to the Gibraltar Strait. Finally, we will show the characteristic modes of variability of the Mediterranean THC resulting from a long numerical simulation under fixed seasonal forcing.
The Atlantic and Mediterranean Sea as Connected Systems
285
5.2.1. Description of Numerical Model The model used is the Modular Ocean Model (MOM, Pacanowski et al., 1991) adjusted to properly simulate the climatological Mediterranean circulation (Artale et al., 2002b; Rupolo et al., 2003a) under different types of air–sea boundary conditions (Pisacane et al., 2006). The model is 1/4 1/4 resolution and has 19 vertical levels with finer resolution towards the surface and coarser resolution at depth. It uses a novel space–time-dependent tracer diffusion scheme, which has proved to be efficient for the numerical simulation of key features of the circulation inside the Mediterranean Sea (Rupolo et al., 2003b). The model Sea Surface Salinity (SSS) is relaxed to monthly mean field climatology of the MODB MED5 (Brasseur et al., 1996). The high-frequency variability of the heat and momentum budget is retained by employing daily surface wind stress of the European Centre Mid-range Weather Forecast (ECMWF) and by relaxing the surface temperature to daily Sea Surface Temperature (SST) values (D’Ortenzio et al., 2000). A wind-dependent relaxation time-scale is assumed (Artale et al., 2002b). The year 1988 has been chosen because the SST is very close to climatological values (see for instance, Marullo et al., 1999) in all the basin except in the Adriatic Sea where the winter SST is too warm (e.g. Gacic et al., 1997). Consequently, a negative bias of about one degree was added to the satellite SST field in February over the Adriatic. These surface heat and momentum forcing, already used by Artale et al. (2002b), produce a reliable description of the principal features of the Mediterranean thermohaline circulation.
5.2.2. Zonal and Meridional Cells All the Mediterranean sub-basins are characterized by water mass formation processes, i.e. intense diabatic processes, and the circulation of the deep and intermediate water masses produced is strongly influenced by topographic constraints and mixing. The main thermohaline cell (order of magnitude 1 Sv) is zonally oriented and primarily driven by evaporation. It is an ‘‘open’’ cell in the sense that both the upper branch and the lower returning flow pass through the Gibraltar Strait. The upper branch is composed of Atlantic Water that propagates through the western and eastern sub-basins, becoming Modified Atlantic Water (MAW) through air–sea interaction and mixing processes. The lower return flow is constituted by LIW whose main formation sites are in the Levantine sub-basin. The LIW is characterized by a relative maximum of salinity, between 200 and 500 m in depth, which can reach values higher than 39.0 psu, which is progressively diluted in its westward spreading (Wu¨st, 1961; Robinson et al., 2001).
286 Mediterranean Climate Variability
Despite this simplified scheme, the detailed circulation in the Mediterranean Sea remains rather complex. In fact, the cold temperatures and strong winds in the northern part of the basin allow for the formation of dense waters also in the Gulf of Lions and in the Northern and Southern Adriatic. These water masses are typically colder, fresher and denser than the LIW and are observed at depths reaching 2,000 m (the MEDOC group, 1970). Therefore, the shallow sills at Gibraltar and at the Sicily Strait prevent such deep water masses from spreading outside the sub-basin where they are first generated. As a consequence, two distinct, closed meridional thermohaline cells are observed in the WM and in the EM. These cells are composed of a northward (in the WM) and north-westward (in the EM) upper branch of both surface and intermediate waters, and a compensating southward/south-eastward lower branch of relatively colder and fresher water (Malanotte-Rizzoli et al., 1997). The western deep cell is mainly driven by the meteorological perturbations peculiar to the North Atlantic systems (Hurrell, 1995) while the Eastern deep cell is driven by the winds coming from the Eastern Europe and Siberia (Lascaratos et al., 1999). The closure of these overturning cells is primarily due to upwelling, mixing and ‘‘pumping’’ processes in the vicinity of the main straits, namely the Gibraltar Strait (Stommel et al., 1973; Whitehead, 1985; Kinder and Bryden, 1990; Kinder and Parrilla, 1987) and the Sicily Channel (Iudicone et al., 2003; Beranger et al., 2004). The global Mediterranean thermohaline circulation is then the result of the interactions between deep water formation and intermediate water dispersal (Wu and Haines, 1996). In fact the intermediate waters, flowing above the sill depths at the main straits, change their properties by travelling around the basin and mixing with water formed in the meridional/zonal cells. Therefore, the MOW, though essentially an intermediate water mass type, is the net result of a complex set of processes involving several water masses (Millot et al., 2006), which will be discussed in detail in the next sections.
5.2.3. Mean State and Variability Due to its relatively small dimensions, the Mediterranean Sea promptly responds to atmospheric variability (Demirov and Pinardi, 2002; Stratford and Haines, 2002). It is therefore quite natural to attribute the observed variability in the production and transformation of water masses in the Mediterranean Sea to interannual changes in the atmospheric forcing at the sea–air interface (Mertens and Schott, 1998; Malanotte-Rizzoli et al., 1999; Castellari et al., 2000; Malanotte-Rizzoli et al., the LIWEX Group, 2003, see also Chapter 4 of this book). On the other hand, the hypothesis that internal feedbacks between the rate of deep water formation and the strength of the overturning circulation
The Atlantic and Mediterranean Sea as Connected Systems
287
might constitute a relevant source of variability in oceanic circulation, independent of changes in the atmospheric forcing, has been proved to be correct (Myers and Haines, 2000). Since the beginning of the 1990s field observations have documented a change in the dispersal of water masses in the EM (Roether et al., 1996; Klein et al., 1999; Lascaratos et al., 1999; Malanotte-Rizzoli et al., 1999, among others). Water of South Aegean origin, which used to be observed at intermediate depth, substituted Adriatic deep water, filling the deep layers in the Ionian and Levantine sub-basins. This bottom water is warmer and saltier than the ‘‘old’’ bottom water of Adriatic origin, which was lifted by several hundred meters. Consequently, the Eastern deep layers became warmer and saltier while the intermediate layers became fresher, colder, relatively poor in oxygen content and richer in nutrients. This event is known as the Eastern Mediterranean Transient (See Chapter 4 of this book).
5.2.4. Timescales of Mediterranean Thermohaline Circulation Observations show that the Mediterranean circulation is not stationary. However, an analysis of the steady circulation regime obtained under climatological forcing can provide an estimate for the characteristic propagation time of water masses from the EM to the Atlantic Ocean. In particular, we have applied Lagrangian diagnostics to the output of the model described in the introduction of this chapter. The methodology for Lagrangian diagnostics has been developed by Do¨o¨s (1995) and Blanke and Raynaud (1997), and has been previously applied in several numerical studies of water mass dispersal (e.g. Blanke et al., 1999, 2001). In this calculation, each particle represents a parcel of water mass of fixed volume. The transport between two sections may be computed by counting particles released in the starting section as they cross the end section or re-circulate through the starting section. By storing the hydrographic properties and the arrival times of each particle, it is also possible to quantitatively estimate the water mass transformation processes and the characteristic dispersal time between the two sections. To describe, in the Lagrangian framework, the Mediterranean circulation we consider about 5 105 numerical particles released in a meridional section in the Alboran Sea. The pathway of each particle is first integrated forward in time until it reaches back to its starting section. This forward integration is used to study the upper branch of the THC. By the forward integration we recognize the existence of well separated ‘‘fast’’, ‘‘western’’ and ‘‘global’’ thermohaline cells. These computations give an eastward flow in the Alboran sub-basin of 1.55 Sv. About 25% of this flow (0.38 Sv) re-circulates rapidly near the Alboran section
288 Mediterranean Climate Variability
without significant hydrographic transformation (‘‘fast’’ cell, dashed arrow in Fig. 90). About 0.35 Sv, spread in the WM without crossing the Sicily Straits (black arrow, western cell). Finally, most particles (0.82 Sv, grey arrow in Fig. 90) enter the EM and contribute to the global thermohaline cell. The existence of three well-separated thermohaline cells is confirmed in Fig. 91, by showing the time distribution of the volume transport associated with the particles released in the Alboran Sea. The figure sketches the distribution of the time necessary for a particle to reach back to its starting section in the Alboran Sea. It is worth noting that the ‘‘western’’ and ‘‘global’’ cells, include both the zonal and meridional cells mentioned in Section 5.2.2. This point is made clear by integrating backwards the particles reaching the Alboran section until they first cross the sections in the Adriatic and Aegean (Fig. 92). The backward integration shows that the westward mass transport of 0.82 Sv computed at Sicily Channel is composed of three different sub-surface waters originated from the following sub-basins: Adriatic (0.25 Sv), Levantine (0.53 Sv) and Ionian (0.04 Sv). Obviously, due to the mass conservation constraint, these waters are produced from the transformation, by strong air– sea interaction within EM, of the same surface water mass entering the Sicily Channel (see Fig. 90).
Global cell Western cell Fast recirculation cell
0.82 Sv 0.35 Sv 0.38 Sv
Figure 90: Schematics of the three zonal cells described in the text.
The Atlantic and Mediterranean Sea as Connected Systems
289
Figure 91: Time distribution of the mass transport associated with the particles reaching the Alboran section after having been initially released in the Alboran section: total transport (thick solid line); transport associated with the fast cell (dashed line), the western cell (thin solid line) and the global cell (dotted line).
P1=SIC-->TYR-->SARD-->ALB= 0.48 Sv P2=SIC-->TYR-->NW-->ALB= 0.15 Sv P3=SIC-->TYR-->SARD-->NW-->ALB= 0.19 Sv
Figure 92: Lower branch of the Mediterranean thermohaline circulation. See text for details. The numbers indicate mass transport expressed in Sv. Particles are released in the Alboran Sea and integrated forward and backward in time in two Lagrangian experiments. In the first experiment, particles are integrated forward in time till they reach again the starting section in the Alboran Sea (ALB). Contrastingly, in the experiment in which particles are integrated backward in time, particles are stopped when they reach the ‘‘ADR’’ and ‘‘LEV’’ section in the Eastern basin.
290 Mediterranean Climate Variability
By considering the dispersal of the returning lower branch in the WM, we find that all the water masses flowing westward across the Sicily Strait enter the Tyrrhenian Sea (Astraldi et al., 2002). Downstream of Sicily, the model simulation suggests the existence of three different paths in the intermediate and deep layers. About 80% (0.67 Sv) re-circulates anticlockwise in the Tyrrhenian Sea and then flows through the Sardinia Channel, where a further bifurcation is observed with about 0.48 Sv (black arrow and P1 in Fig. 92) flowing directly toward the Alboran Sea and 0.19 Sv (blue arrow, P3 in Fig. 92) flowing north toward the North Western Mediterranean Sea (NWM) before flowing toward the Gibraltar Strait. The remaining 18% (0.15 Sv, red arrow and P2 in the same figure) exits the Tyrrhenian Sea from the Corsica Channel and flows to the NWM before reaching the Strait of Gibraltar. In Fig. 93, we show the time distribution of the volume transport associated with the particles reaching the Alboran Sea from the Sicily Strait for the paths P1, P2 and P3. Since the distribution of the arrival times is characterized by very long tails it is interesting to consider different mean arrival times for each pathway (Table 6). The mean time ranges from 17 years for the fastest path P1 to 73 years for the slowest path P2, the latter including the dispersal of the deep water formed in the NWM.
Figure 93: Time distribution of the mass transport associated with particles reaching the Alboran Sea from the Sicily Strait of the total flow (thick solid line) and the paths P1 (thin solid line), P2 (dashed line) and P3 (dotted line). See Section 5.2.4 for a definition of the three paths.
The Atlantic and Mediterranean Sea as Connected Systems
291
Table 6: Mode, median and mean times of the arrival times for the three paths of the flow from the Sicily Strait to the Alboran Sea defined in the text. Times are expressed in years.
P1 P2 P3
Tmode
Tmedian
Tmean
5 16 7
10 48 19
17 73 37
By defining the age of a water mass as the time necessary to go from the Sicily Strait to the Alboran Sea, we can also quantify the percentage of water mass younger than a threshold value through the computation of the ratio between the cumulative function of the distribution of the arrival times of a given path and that of the total path. Figure 94 shows that very young water mainly consists of water flowing along path P1, while for large times (old waters) the relative weight of slower paths continuously increases, reaching, for path P2, the asymptotic value after about 150 years. These diagnostics indicate that a T/S anomaly entering the WM from Sicily Strait is scattered to the three different paths characterized by very different
Figure 94: Ratio between the cumulative distribution function of arrival times for the three paths defined in the text and that of the total flow. The continuous, dashed and dotted lines refer to the paths P1, P2 and P3.
292 Mediterranean Climate Variability
propagation times. On the other hand, the same results indicate that a change in the hydrological properties of the slowest path P2, influenced by strong air–sea interactions and deep water formation in the NWM, may impact after large times on the characteristics of the MOW. Finally, we show in Fig. 95 the mean T/S properties in the different sections defined in Fig. 92. The hydrological properties of the intermediate flow, observed in the Sicily Channel (SIC in Fig. 95), is the result of a mixing between the following water masses: the cold and fresh Adriatic Deep Water (ADW, ADR in Fig. 95), the warm and salty LIW (ADW, LEV in Fig. 95) and surface less-dense (warmer and fresher) water mass (not shown in Fig. 95). In other words, these numerical results indicate that in the Ionian Basin there occur diabatic processes in which the mixing between the eastern deep water and shallower waters give rise to a westward flow in the Sicily Strait that is less dense than the original ADW and LIW. In the WM, as observed for instance by Wu¨st (1961), it is possible to see that in the first part of the lower branch of the cell (namely in the TYR and SARD section in Fig. 92) the hydrological properties of the flow remain rather constant (Millot, 1999). In contrast, the T/S diagram shows that the hydrological properties of the flow in the Alborean Sea (ALB in Fig. 95) lies in between those of the T/S values observed in the SARD and NW sections, confirming the importance of mixing processes between the Intermediate Water (IW) coming from the EM with the water mass types originating in the Northwestern sub-basin.
Figure 95: Mean T/S values of the particles crossing the different sections defined in Fig. 92.
The Atlantic and Mediterranean Sea as Connected Systems
293
To summarize, the hydrological properties of the Alboran sub-surface water, that afterward forms the MOW, depend on the interaction between LIW, ADW and the MAW in the EM and between the IW of eastern origin and the cold and denser water formed in the north-western basin in the WM. The mixing with this colder and fresher water changes the salinity and temperature by about 0.2 psu and 0.5 C. Because the variability of deep water formation and therefore the freshening of the shallow depth depends on winter meteorological condition, any change in the atmospheric circulation, and in particular the appearance of extreme meteorological events, may affect the hydrological properties of the MOW. Moreover, the previous discussion on the characteristic dispersal times of the lower branch of the Mediterranean Thermohaline cell indicates that a change in the eastern IW source can propagate towards Gibraltar rather rapidly (see the typical arrival times of the direct path P1 from the Sicily to the Gibraltar Strait in Table 6), while any change in the NWM water mass formation is diluted over many years and may only have a major influence after long times. Obviously these results have to be considered only as an indication of the possible phenomenology of a circulation change, since they are obtained with a numerical model forced with a perpetual-year atmospheric forcing field.
5.2.5. Decadal Oscillations in the Mediterranean Sea The conjecture that advective–convective internal feedbacks might constitute a source of variability in oceanic circulation, independent of the external forcing, has been proved by several studies (Lenderink and Haarsma, 1994; Yin and Sarachik, 1995). Thorough understanding of such mechanisms inside the Mediterranean might prove crucial, not only in view of the potential impact of climatic variability on such a densely populated area, but also as it could provide material for climate studies at global scales. In fact, the Mediterranean Sea may be considered as a test basin, where processes are expected to occur on timescales which enable both close experimental monitoring and affordable numerical simulations. To test this idea, we performed a long numerical simulation with the model configuration described in Section 5.2.1 (about 1200 years; Pisacane et al., 2006). The circulation exhibited remarkable variability in local hydrological properties, with oscillations over decadal timescales both in salinity and temperature. Moreover, periodically enhanced convection was observed in the Gulf of Lions. The use of external perpetual-year forcing eliminates the possibility that the observed fluctuations can be induced by atmospheric interannual variability. Therefore, the observed fluctuations must be generated by internal (say oceanonly) convective–advective feedback mechanisms. In a climatological perspective, it is desirable to deal with integral quantities whose time evolution and spatial distribution are descriptive of the large-scale variability, whereas local indicators
294 Mediterranean Climate Variability
would mask important information with high-frequency noise. In the EM, zonal and meridional transports provide valuable synthetic information on the dynamics at basin scales and on the local variability. A decomposition of the transport fields into EOFs identified the main patterns of variability (Fig. 96). A
B
Figure 96: First mode and second mode of the EOF decomposition of the Meridional Transport (expressed in Sv) in the Eastern Basin.
The Atlantic and Mediterranean Sea as Connected Systems
295
The first two modes explain approximately 60% of the variance, and mode 2 in particular supports the hypothesis of a negative correlation between processes in the Aegean and in the Adriatic Sea, which is intriguing in view of the nature of the EMT. This mode has a positive lobe inside the Adriatic Sea which extends southward into the Ionian sub-basin, while a deeper negative lobe is located at a latitude that corresponds to the Aegean sub-basin. A possible interpretation of this result is that the variability of the meridional overturning cell is not solely related to the variability of deep water formation in the Adriatic Sea, but is affected by processes that are located elsewhere in the EM. The spectrum of the Principal Component associated to the first EOF (Fig. 97) shows a well-defined peak of variability corresponding to a period of about 7 years, consistent with observations in local series of temperature and salinity. The EOF analysis carries no significant dynamical information. It only serves to isolate coherent correlation structures over the timescales of interest. A question arises as to whether the variability in the mean transport is a signature of the weakening/strengthening of the circulation without changing its structure, or if the circulation has different equilibria in which the competing roles of the Aegean and the Adriatic Seas as sites of deep water formation are altered. The existence of multiple equilibria of the EM has been
Figure 97: Power spectrum of the principal component corresponding to the first EOF shown in Fig. 96. The abscissa axis represents time, expressed in years.
296 Mediterranean Climate Variability
recently investigated by Ashkenazy and Stone (2005) using a 3-box model to study the salinity and heat fluxes changes in the Adriatic–Ionian–Aegean system. Even though the model represents a simplification of the EM circulation, several sensitivity experiments show that the current thermohaline circulation of EM might be drastically changed. In particular the role of the Adriatic is crucial to define the stability of the EM thermohaline circulation, being the dominant source region both for salinity and temperature in comparison with the Aegean, which plays the same role only for temperature. In our simulation two different patterns can be detected in the time series of monthly mean meridional transport fields in the EM, which are shown in Fig. 98. In the following, these two patterns will be indicated as pattern A (panel A of Fig. 98, strong intermediate meridional cell) and pattern B (panel B, weak meridional cell). Furthermore, the meridional and zonal overturning are correlated, as shown in Fig. 99, indicating that the whole EM undergoes a cyclic variation in the intensity of its circulation, which can be related to the alternative predominance of the Adriatic or the Aegean Sea in the production of dense water. In order to underline the alternative role of these two basin (with the dominance of the Adriatic) we construct, for each month, the time series of the differences (DD) between the density of the water at the sill level in the Aegean Sea (SAW) and the density of the outflow (i.e. the ADW) at Otranto Strait (i.e. DD ¼ SAW ADW), and correlate it with the time series of the maximum overturning. Figure 100 shows that DD and the meridional overturning circulation are always anticorrelated, and that the anti-correlation is maximum in correspondence of deep water formation events. Such an anticorrelation shows that a predominance of SAW (positive values of DD) corresponds to a weaker cell (pattern B). Even though the absolute value of the correlation minimum is fairly low, the result is encouraging, considering that local density values are correlated with a basin-scale variable that keeps track of remote dynamics. Although the statistical relevance of these correlations should be supported by a close analysis of the 3D dynamics, the relevance of natural oscillations in phenomena such as the EMT deserves more attention. From such a perspective, the role of the atmospheric forcing would not be diminished, but it would need to be reconsidered in relation to the state of the basin, onto which it is impulsively acting. In particular, it may be argued that only if the Aegean and the Adriatic Waters have already altered their relative weights due to natural variability, will the forcing exerted by the atmosphere produce a really significant effect. The interpretation of the EMT as a purely atmosphere-driven event should be reviewed, and the role of internal variability and advection–convection feedbacks, should be reconsidered.
The Atlantic and Mediterranean Sea as Connected Systems
297
A
B
Figure 98: Monthly meridional transport (units Sv) in the Eastern Basin for the period 15 Jul/14 Aug in two selected years: 351 (A) and 359 (B). The impact of the competing roles of the Adriatic and Aegean Sea on the variability and hydrological characteristics of the water masses exported from the eastern to the western sub-basin is still to be understood. Local dynamical variations occur, such as in deep water formation in the Gulf of Lions, where
298 Mediterranean Climate Variability
Figure 99: Scatter diagram of the meridional vs. zonal overturning. The correlation coefficient is r ¼ 0.5.
convective events can sometimes be enhanced, showing a periodicity over decadal timescales. However, in the WM, the decomposition into EOFs fails to capture a high degree of variability in the circulation, which might have been expected as a consequence of the oscillations in the LIW properties detected in local time series. However this is not surprising in view of the WM processes that are liable to be affected by LIW variability. In particular, winter convection in the Gulf of Lions is a phenomenon more localized both in space and time if compared with the larger spatial and temporal scales involved in the water masses transformation within the EM. Therefore, it is a good example of the differences in the dynamics of the western and the eastern sub-basins. However it is hazardous to conclude that any variability in the WM convection is due to the preconditioning effect of LIW with variable temperature and salinity. In fact, any correlation analysis between local time-series in the Gulf of Lions and in the EM is undermined by the fact that the abrupt onset of convection is dependent on threshold values of water properties, and also on local factors that are independent of the signal
The Atlantic and Mediterranean Sea as Connected Systems
299
Figure 100: Seasonal variation of the correlation coefficient between the density difference DD and the maximum meridional overturning in the eastern basin. transferred through the Sicily Channel by the advected LIW. Moreover, the climatological signal might be concealed by high-frequency noise. Therefore, the analysis of the WM demands the development of suitable statistical diagnostics that are capable of relating phenomena occurring over distinct space and timescales. The question of teleconnections cannot however be overlooked, and the impact of the variability in LIW properties on the preconditioning of winter convection in the Gulf of Lions should be more thoroughly addressed.
5.3. The Strait of Gibraltar: A Gate to the Atlantic The mechanisms that determine the budget of fresh water imported into the Mediterranean Sea from the Atlantic Ocean and the corresponding mechanisms generating the outflow of Mediterranean water into the North Atlantic are key issues to understand how the Mediterranean–North Atlantic system is working.
300 Mediterranean Climate Variability
The AW is the source of fresh water available for convection and vertical water masses transformation processes. The Strait of Gibraltar is the first and perhaps the most important limiting factor for the availability and for transformation of this water mass within the three-dimensional Mediterranean circulation. Therefore, an important role is played by local phenomena such as hydraulic control and mixing processes within the strait that in turn depend on density and pressure differences between the Atlantic and Mediterranean Sea (Nof, 1991; Hopkins, 1999). Due to the steepness of the strait bathymetry over a relatively short distance, the flow of Atlantic and Mediterranean waters rapidly varies and a local hydraulic jump is usually present. The entrainment and mixing between the Atlantic upper layer and the Mediterranean lower layer is responsible for the creation of a third interfacial layer as observed by Bray et al. (1995) and simulated by Sannino et al. (2004). Local phenomena occurring at the strait can also have an impact on larger scales and may explain the differences between the trends in temperature and salinity observed in the outflow (Millot et al., 2006) and those observed in the Western Mediterranean Deep Water (Rixen et al., 2005), the difference being one order of magnitude lower. In fact, in the last decade (1994–2004), a large warming (0.3 C) and salinity increase (0.06) has been observed at the sill at Gibraltar (Millot et al., 2006). From 1955 to 1993, the trends are 0.1 C and 0.02/decade in the 10 10 zone west of Gibraltar (Potter and Lozier, 2004) and of almost the same magnitude even west of the mid-Atlantic Ridge (Curry et al., 2003). According to Vargas et al., 2002, there is experimental evidence of a positive trend both in temperature and salinity during years 1995–2002 ( 0.5 C/decade and 0.1 psu per decade, respectively). These changes are much larger than the trends of 0.01 C/decade observed in most waters at a global scale (Levitus et al., 2000). Specifically, WMDW has warmed since the 1960s (Rixen et al., 2005), but its linear trends of 0.035 C and 0.01 psu per decade are much less than those observed in the Mediterranean outflow west of Gibraltar. Furthermore, as a consequence of specific hydraulic conditions, for example when the flow is in maximal regime, the Gibraltar Strait may contribute to isolating the Mediterranean Sea from the rest of the world ocean (Tsimplis and Baker, 2000). We use the results of a high-resolution numerical model (Sannino et al., 2002a, 2004) to assess the dependence of mass transport on the salinity difference between Atlantic and Mediterranean waters, and on the tidal flow. The model used for this study is the three-dimensional, sigma coordinate Princeton Ocean Model (POM) of Blumberg and Mellor (1987) in a very high-resolution configuration (x, y < 500 m within the strait) which allows for a detailed description of all dominant topographic features within the strait. It is well known that the surface inflow is determined by the Atlantic low-salinity water (S ¼ 36.5), and the outflow at depth by the Mediterranean colder high-salinity
The Atlantic and Mediterranean Sea as Connected Systems
301
water (S > 38) (Lacombe and Richez, 1982; La Violette, 1990), with an average value of S of about 1.5 psu. Changes in temperature and salinity of the inflowing waters can be attributed to hydrological changes in the North Atlantic (La Violette, 1990). Inflowing Atlantic water forms a low-density surface layer of 100–200 m thickness, which becomes the MAW (La Violette, 1994). Changes in temperature and salinity of the Mediterranean outflow are due to the variability of the deep and intermediate waters within the entire Mediterranean Sea, as discussed in several sections of this book (see in particular Chapter 4 and Section 5.2). Our simulation shows a hydraulic jump west of the western sill of the Strait (Camarinal Sill), whereas there is no similar phenomenon in the eastern part of the Strait (Tarifa Narrows). In particular, at Camarinal Sill the mean cross-strait composite Froude number is found to be in the range 1.2–1.5, whereas at the Tarifa Narrows it is between 0.4 and 0.8 (Sannino et al., 2002a). This result allows us to conclude that the mean circulation simulated by the model is in a sub-maximal regime. By performing other sensitivity experiments with different initial conditions (i.e. S ¼ 2.0 psu, S ¼ 2.2 psu and S ¼ 2.5 psu, corresponding to increased density differences) our model predicts a transport in both layers of 0.86, 0.89 and 0.93 Sv, respectively (Fig. 101). The simulations have shown that when S increases the upper layer Froude number also increases; moreover its range of action extends along the cross section located east of Tarifa (Fig. 102). In particular, for values of S > 2 the mean cross-strait composite Froude number is greater than 1, both at the Camarinal Sill and at the Tarifa Narrows,
Figure 101: Predicted transport through the strait for the three new experiments (A) S ¼ 2.0 psu, (B) S ¼ 2.2 psu and (C) S ¼ 2.5 psu (see Fig. 102; Sannino et al., 2002b)
Spain
Morocco A
Spain
Morocco
B
Spain
Morocco
C
0.0
0.2
0.4
0.6
0.8
1.0
1.2
1.4
1.6
1.8
2.0
Froude Number Figure 102: Predicted three-layer composite Froude number for the three new experiments (A) S ¼ 2.0, (B) S ¼ 2.2 and (C) S ¼ 2.5. Contour lines represents (thin) the bathymetry, (bold) limit the region where G2 > 1.
The Atlantic and Mediterranean Sea as Connected Systems
303
i.e. the flow is in a maximal regime. The result allows us to conclude that the mean circulation simulated by the model is in a sub-maximal regime for S < 2.0 psu and in a maximal regime for S > 2.0 psu. Therefore, small changes in the hydrological characteristics of water masses on both sides of the strait may imply a sudden qualitative change in flow regime observed at the strait. Considering for example a climate change scenario involving a positive trend of the density difference between the Mediterranean sub-surface waters and the Atlantic surface water, we can predict a partial isolation of the Mediterranean Sea from the rest of the world Oceans (Tsimplis and Baker, 2000). That is possible because the non-linearity of the interaction between largescale ocean climate change and local strait phenomena may sustain an abrupt change in the salt/freshwater transport between the Mediterranean and the Atlantic. Finally, the presence of tides increases the mean transport, and the hydraulic behaviour within the strait becomes much more complex (Brandt et al., 2004; Sannino et al., 2004).
5.4. Spreading of Mediterranean Outflow Water in the North Atlantic The water mass transformation and exchange processes described in the previous section, lead to the formation of a characteristic tongue of Mediterranean Outflow Water within the North Atlantic. The presence of this tongue is a persistent and robust feature. However, as discussed in the introduction of this chapter, the quantitative details concerning the spreading of MOW in the North Atlantic are still uncertain. There is also observational evidence of substantial changes to this picture. Curry et al. (2003) showed an increase of both temperature and salinity along a section in the western Atlantic Ocean, which is directly affected by the presence of MOW. By ‘‘directly’’ we mean that MOW is advected to those regions by well defined and rather stable flows. In the eastern part of the basin significant changes are observed even on relatively short timescales. In Fig. 103 we show the salinity observed along a section at 36 N, which has been repeated over at least 15 years from 1976 to 1989 (see panel C of the same figure). The salinity field (panel A) shows the typical signature of MOW with a positive salinity anomaly at depths ranging from 700 to 1500 m. The difference between the periods 1981–1985 minus 1976–1980, (panel B of the same figure), suggests a lift up of the whole layer of MOW, at least in this small region. Analogous changes are observed in the temperature field, which is not shown. In correspondence to these changes, the analysis of Potter and Lozier (2004) shows a saltening and warming of the layer of MOW during the same years. The timescales (around 5 years) over which these changes occur, are interestingly
304 Mediterranean Climate Variability
Figure 103: Salinity section for the periods 1976–1980 (A) and the difference 1980–1985 minus 1976–1980 (B) along the section shown in panel (C). similar to the timescale of natural variability suggested by the numerical simulation discussed in Section 5.2.5. The changes shown in Fig. 103 suggest a possible mechanism for the variability of MOW, originating from the water mass characteristics which are
The Atlantic and Mediterranean Sea as Connected Systems
305
produced inside the Mediterranean. In the following section, we review present knowledge about the dynamics of MOW. In particular, we focus on those processes occurring during the first stages of the spreading of MOW in the North Atlantic that are most likely to be affected by changes occurring inside the Mediterranean Sea. We will also discuss some results concerning the possible impact of the spreading MOW on the pattern of circulation and variability of the Atlantic Meridional Overturning Circulation (MOC), and therefore on climate.
5.4.1. Dynamics of Mediterranean Outflow Water Two main paths are observed for the spreading of MOW in the North Atlantic at intermediate depth. One is the Westward Tongue (WT), extending from the Gulf of Cadiz towards the central North Atlantic at mid-latitudes. This path is characterized by the presence of stable sub-surface eddies containing Mediterranean water, so called meddies, that are advected by the large-scale wind-driven circulation (Nof, 1982; Hogg and Stommel, 1990) and play the role of heat and salt sources to the Mid-North Atlantic. Another path is the Mediterranean Undercurrent (MU), off the Iberian Peninsula, which is in geostrophic balance with the ocean interior and leads to slow mixing with the surrounding water masses (Spall, 1999). A quantitative description of the two branches of flow is given by Sparrow et al. (2002) who analysed recent float data and found, north of 36 N, a background northward flow, with an average velocity of 1.8 0.6 cm/s, and peaks of 10.1 3.7 cm/s off the coast of Portugal. South of 36 N they found a weaker background flow (about 0.12 cm/s), and strong mesoscale activity with peak energy as high as 89 cm2/s2. Meddies The strong mesoscale activity of the WT was first reported by McDowell and Rossby (1978) who introduced the term meddy, to indicate sub-mesoscale coherent vortices containing water of Mediterranean origin. Armi and Zenk (1984) described the kinematics and hydrology of three individual meddies. These were coherent structures about 50–140 km wide and were observed to be stable over periods of years. Richardson et al. (2000) observed meddies as old as five years and estimated an average lifetime of 1.7 year. Along with a meddy formation rate of 17/year, estimated by Bower et al. (1997), who examined the initial growth of meddies off the coast of Portugal, this gives an expected number of 29 meddies co-existing in the North-Atlantic. However only 14 could be observed in 1994. Among these, 11 where observed during the month of February. Although the strong mesoscale activity of the WT has been investigated in great detail, very different estimates of the role and relative contribution of meddies to
306 Mediterranean Climate Variability
the total westward transport of salt exist. Richardson et al. (1989) suggests a contribution of about 25% of the total transport of MOW based on statistics of observed meddies; Arhan et al. (1994) estimates a 55% contribution of meddies; Maze´ et al. (1997) attribute to the detachment of meddies from the slope undercurrent almost all of the westward advection of the salinity anomaly (100%). Mediterranean Undercurrent The Mediterranean Undercurrent path has also been the object of much attention but its path is still very uncertain. Two similar cruises, Bord-Est 2 and Bord-Est 3, conducted in 1988 and 1989, analysed by Maze` et al. (1997) and Arhan et al. (1994), produced quite different pictures of the Eastern North Atlantic. In the first case (Bord-Est 2), a southward surface flow of 2–3 Sv off the coast of Portugal was inferred from hydrographic sections. Similar results are reported by Krauss and Ka¨se (1984), Roether and Fuchs (1988), and Paillet and Mercer (1997). In the second case (Bord-Est 3), a total northward transport of 9–12 Sv was observed, with a well-defined maximum in the MOW layer. A northward flow off the coast of Portugal was also observed by Haynes and Barton (1990) and by Rios et al. (1992). Many different factors, such as the interactions with regional topographic features, the superposition of water masses, the strong mesoscale activity and even the seasonality of the Azores High, may seriously affect all indirect measurements of both the WT and MU paths in this region. Further north, the MOW can be clearly tracked at 46 N in the central Bay of Biscay (van Aken, 2000). Two different scenarios have been proposed for the way the MOW can affect the MOC in the Atlantic Ocean. The first one was proposed by Reid (1979) in which the MOW is directly advected to the Greenland– Iceland–Norwegian (GIN) Sea on intermediate layer density horizons. Iorga and Lozier (1999) also support the so-called deep source hypothesis, whereby a substantial northward transport of MOW, through the Rockall Channel, may directly affect the preconditioning in the GIN Sea to overturning. A second scenario was originally proposed by Lozier et al. (1995) in which the MOW progressively ‘‘peels off’’ interacting with the North Atlantic Current and modifying its water properties. More recently, the direct advection/deep source hypothesis has been severely criticized on the basis of new hydrographic observations by McCartney and Mauritzen (2001), who strongly support the alternative shallow-source hypothesis, i.e. the scenario proposed by Lozier et al. (1995). In this case the warm, salty inflow into the Nordic Seas is attributed to the northward branching of the North Atlantic Current. Bower et al. (2002) have recently attempted a direct measurement of the middepth circulation of the eastern North Atlantic. They shed new light on the MU pathway in the North Atlantic observing that the MOW is diverted into the
The Atlantic and Mediterranean Sea as Connected Systems
307
ocean interior from the Bay of Biscay and ‘‘probably penetrates north of 52 N by mixing with the North Atlantic Current in the complex area southwest of the Porcupine Bank (49N; 16W)’’. Thus, relatively little water of Mediterranean origin is found inside the Rockall Trough (53N; 15W), either at the surface or at depth. Rather, almost all of the North Atlantic Current crosses the ridge and flows into the Icelandic Sea. To summarize, different pictures of the spreading of MOW in the North Atlantic have been constructed from the observations and no definitive agreement has been reached yet about any of them. Most importantly, very little is known about the variability of these pathways.
5.4.2. Mediterranean Outflow Water and Stability and Variability of Meridional Overturning Circulation Since the Atlantic Ocean contributes a substantial part of the poleward heat transport from the tropics, all major climatic shifts on all timescales have been related to major changes in the Atlantic Meridional Overturning Circulation. On centennial and millennia timescales, the onset of different convective regimes in the North Atlantic has been deduced from proxy data (e.g. sediment cores) and it is probably important for the structuring and rapidity of the prevailing climate changes (Dokken and Jansen, 1999; McManus et al., 2004). On paleoclimatic timescales, both observational and modelling evidence support the hypothesis that the transition from the Last Glacial Maximum to the Holocene has been related to a major reorganization of the overturning circulation of the North Atlantic (Boyle and Keigwin, 1987; Sarnthein et al., 1994; Seidov and Haupt, 1997). Moreover, modelling studies suggest that under glacial conditions the meridional overturning circulation may be characterized by different stability properties than under present-day conditions (Ganopolski and Rahmstorf, 2001). It is generally proposed that two fundamental processes are in opposition to each other, leading to instability and to rapid transitions between the different modes of operation of the overturning circulation, namely the advective and convective feedback processes (Rahmstorf, 1996). The former is related to the lateral advection of warm and salty waters towards areas of deep water formation. The excess of heat is rapidly lost through enhanced exchanges with the atmosphere while the salt content is retained. In this way, the instability of the water column is enhanced and, by mass continuity, a positive feedback on the circulation is created. The convective feedback relies on a similar exchange mechanism but operates in the vertical during the onset of deep convection, and is related the localization of convective sites (Lenderink and Haarsma, 1994). Tziperman (2000) illustrates how these two mechanisms may interact and
308 Mediterranean Climate Variability
generate the instability of a given circulation regime and the formation of a new equilibrium. The impact of the MOW on the strength of the overturning circulation of the North Atlantic was observed to be fairly weak (modelling result) by Rahmstorf (1998). On the other hand, Hecht et al. (1997) achieved a stable MOC only when a realistic MOW was present. Recently Artale et al. (2002a) have proposed a synthesis between these two rather different points of view, by showing that the Mediterranean Outflow has only a slight influence on the strength of the MOC but a considerable impact on the preference for certain patterns of circulation and their stability. In Fig. 104 we show the mechanism which is responsible for this behaviour.
Figure 104: The enhancement of advective feedback in the presence of MOW, see Section 5.4.2 for details (from Artale et al., 2002a).
The Atlantic and Mediterranean Sea as Connected Systems
309
We integrated the two-dimensional Boussinesq equations under mixed boundary conditions, with a temperature profile and freshwater flux symmetric around the equator. The differences between two similar overturning regimes, which differ by the presence of an intermediate depth anomaly mimicking the presence of MOW, shows that the advective feedback described above is favoured by the presence of MOW. A stronger temperature gradient in the most northern region shows up in the presence of MOW. This allows a partial damping of the temperature anomaly, while the salinity anomaly is more easily advected northwards, thus determining a net positive density anomaly near the sinking region. The increased surface density then enhances the MOC, as shown by the stream function anomaly. Different issues are raised in the context of climate variability on shorter timescales. At interdecadal timescales, oscillatory modes of the Atlantic MOC exist, that may be easily explained in terms of linear dynamics around a thermally direct equilibrium (Griffies and Tziperman, 1995). For example, the variability observed in a coupled model experiment by Delworth et al. (1993) has been explained in terms of this oscillatory mode. Griffies and Tziperman (1995) suggested that the variability is maintained by the presence of a stochastic atmospheric forcing which excites the damped oscillatory mode of the system. Following these ideas, it is possible to attribute a large portion of the variability of the MOC to the existence of a positive internal advective feedback, which is possible only when salinity gradients can be maintained in the ocean interior, due to the presence of the MOW source at intermediate level. Figure 105 show power spectra obtained by integrating a simple Stommel-type-box model under stochastic surface forcing (Stommel, 1961). The model includes an intermediate layer which is forced independently from what happens at the surface. The intermediate level forcing is tuned to represent the spreading of MOW in the North Atlantic. In Fig. 105 we can distinguish three different scenarios. In one case (NoMed) there is no Mediterranean Outflow in the intermediate layer. In the scenario A, which has reduced variability, the Mediterranean Outflow mixes directly with newly formed water masses at high latitudes. Finally, in the scenario B, we consider the situation in which the Mediterranean Outflow upwells somewhere in the North Atlantic before entering the main water mass formation sites, in the Labrador Sea or in the Greenland–Iceland–Norwegian Sea. Therefore, depending on the processes that determine the spreading of MOW in the North Atlantic, the effect on the variability of the MOC in the Atlantic can be very different. This raises the importance of a correct description of the Mediterranean Outflow in the oceanic component of climate models, which we discuss further in the following section.
310 Mediterranean Climate Variability
0.1
NoMed A B
2
Sv / cpy
0.01
0.001
0.0001
1e-05 0.01
0.1 f [cycles per year]
Figure 105: Power spectrum of the meridional overturning circulation, with different scenarios of the spreading of MOW in the North Atlantic. See Section 5.4.2 for details.
5.4.3. Modelling of Mediterranean Outflow Water Simplified Models Much effort has been devoted to the understanding of the mechanisms that contribute to the shaping of the Mediterranean Outflow. Motivated by the observations of Arhan (1987), Tziperman (1987) explained the shaping of MOW in terms of Rossby waves originated from the thickening of intermediate depth isopycnal layers at the eastern boundary. Spall (1994) analysed the linear instability of the meridional flow in the eastern North Atlantic, and found that baroclinic instability could be responsible for a significant fraction of the westward transport. The role of ventilation has been considered by Stephens and Marshall (1999) who stressed the crucial role of the wind forcing in shaping the salinity tongue in the central Atlantic, while meddies were described as a distributed source of salt in the interior of the ocean. One of the most serious limitations to modelling the dynamics of the MOW is the existence of a shallow and narrow communication between the North Atlantic and the Mediterranean Sea at the Strait of Gibraltar. A common
The Atlantic and Mediterranean Sea as Connected Systems
311
approach to overcome this limitation is to restore values of temperature and salinity at the eastern boundary to climatology (Stanev, 1992; Hecht et al., 1997; Rahmstorf, 1998; New et al., 2001). However, Gerdes et al. (1999) found that realistic WT and MU can be obtained only by the imposition of an actual inflow at the eastern boundary. This result has serious consequences for the design of an ocean GCM suitable for climate studies. General Circulation Model Studies Several GCM studies of the spreading of the Mediterranean Outflow in the North Atlantic can be found in the literature. Hecht et al. (1997) first noted the importance of a proper representation of the outflow for the stability of the MOC in the North Atlantic. However, relying on different model configurations and parameters, Rahmstorf (1998) stressed that only minor changes in the intensity of the MOC occur if the Mediterranean Outflow is switched off. This result is confirmed in a more recent study by Chan and Motoi (2003). No estimates of the potential impact of future changes exist to date. Numerical simulations of global warming scenarios, suggest that the Mediterranean Outflow is likely to become warmer and saltier, and probably shallower, than today during the next century (Thorpe and Bigg, 2000). However, the poor description of the MOW in coupled climate models makes it difficult to have confidence in the current predictions. It proves difficult to correctly describe the descent, adjustment and mixing of the Mediterranean Outflow with the relatively coarse model resolutions employed in climate studies. Recent numerical simulations (Sannino et al., 2004) have shown that tides contribute up to 30% of the total exchange at the strait. They also showed that the depth of the outflow is substantially affected by changes in its salt content, thereby suggesting that the dynamics of the Strait of Gibraltar deserves a more careful treatment. Usually, the spatial resolution employed in climate models is 1 1 or less. This resolution is insufficient to resolve the processes occurring at the Strait of Gibraltar (13 km). Furthermore, the description of the processes occurring in the Gulf of Cadiz, which set the hydrographic characteristics and precondition the dynamics of the outflow in the North Atlantic, proves to be inadequate even in models with finer resolution (Treguier et al., 2003). In these cases, the interplay between mixing processes, entrainment and bottom stress is problematic and requires a special treatment. Typically, the problem encountered with z-level models is the large impact of lateral mixing that produces too shallow overflows. On the other hand, layered models have too low diapycnal mixing, resulting in too deep overflows (Willebrand et al., 2001). In the case of MOW, such deficiencies result from an incorrect description of the dynamical balances, especially in the coarse resolution models employed for climate studies. For example, after integrating to
312 Mediterranean Climate Variability
equilibrium the 2 2 standard configuration of the OPA model, we found that the Mediterranean Undercurrent could not be reproduced, whereas in a higher resolution GCM by Jia, Coward, de Cuevas, Webb and Drijfhout (personal communication), a consistent Mediterranean Undercurrent is found, even though the outflow is about 200 m shallower than observed. Treguier et al. (2003) observe similar deficiencies. As to the MOW spreading in the central north Atlantic, the studies of Stephens and Marshall (1999) and Richardson and Mooney (1975) stress the dominance of advective processes in shaping the Mediterranean Salinity Tongue. This may also be a limiting factor in climate models where a large value of lateral diffusivity is used for computational reasons, adversely affecting the dynamics of tracer diffusion. In Fig. 106, we compare the climatology of Levitus with two different simulations performed with the OPA model (Madec et al., 1998). In Fig. 106B, the equilibrium state of the standard configuration is shown, where diffusion
(A)
Figure 106: Comparison of salinity field on the neutral isopycnal surface
¼ 27.70: the climatology of Levitus (panel A); the spinup of the standard configuration of the OPA model at 2 2 resolution (panel B); the spinup of the OPA model at the same resolution but with the space–time-dependent tracer diffusion scheme described in Rupolo et al. (2003b).
The Atlantic and Mediterranean Sea as Connected Systems
(B)
(C)
Figure 106: Continued.
313
314 Mediterranean Climate Variability
dominates over advection, resulting in an unrealistic salinity tongue. An improvement is observed by employing the same lateral diffusion scheme as in Rupolo et al. (2003b). The corresponding salinity field is shown in Fig. 106C, the shape of the Mediterranean Salinity Tongue is narrower and more consistent with the observations. However the isopycnal that was chosen to represent the spreading of MOW is slightly too deep. This result shows that an improvement in the tracer dispersion process is likely to add important details to the modelling of the spreading of MOW in the North Atlantic.
5.5. Discussion and Future Challenges In this chapter, we have discussed the impact of the Mediterranean Sea on the Atlantic Ocean and how some processes occurring in the Mediterranean Sea, and in particular the production of intermediate waters, provide a paradigm for analogous processes occurring in the world ocean (Lacombe et al., 1985; Bethoux et al., 1999; Artale et al., 2002a). The Mediterranean Sea directly experiences the influence of the North Atlantic, both directly through storms track perturbations (Hurrell, 1995), and indirectly via the Atlantic Water inflow at the Strait of Gibraltar (Robinson et al., 2001). We found that the variability of intermediate water characteristics for the Mediterranean and Atlantic (LIW or MOW) provides the ‘‘fingerprint’’ of the ocean internal variability. In the case of Mediterranean the EM plays the role of the ‘‘engine’’ of its internal thermohaline cell, in which effective transformation of the surface water into intermediate water occurs. Our numerical simulations have also shown that the natural decadal variability of the Mediterranean Sea thermohaline cell is related to the amount of water mass that flows through the Sicily Channel. This water mass is a mixture of the water masses generated inside the Aegean and Adriatic sub-basins. In particular, a ‘‘sea-saw effect’’ between the Aegean and Adriatic Sea is observed, where the reservoir of dense water in the Aegean sub-basin reaches a threshold value (the role of the sills is crucial), above which the Aegean becomes the principal deep water source for the EM. This process then inhibits the production of deep water in the Adriatic sub-basin and vice versa. The complex interaction between external forcing and internal variability in processes may provide the explanation for the Eastern Mediterranean Transient. Changes in the dispersal of water masses in the EM since the beginning of the last decade (1990s) have been documented by several analyses of field observations demonstrating that water of Aegean origin, usually found at intermediate depths, substituted the Adriatic DW, filling the deep layers in the Ionian and Levantine seas.
The Atlantic and Mediterranean Sea as Connected Systems
315
The variability in the production and transformation of water masses in the Mediterranean Sea has been mainly attributed to interannual changes in the strength of the external atmospheric forcing (Roether et al., 1996; Mertens and Schott, 1998; Malanotte-Rizzoli et al., 1999; Samuel et al., 1999; Theocharis et al., 1999; Demirov and Pinardi, 2002; Malanotte-Rizzoli et al., the LIWEX Group, 2003; Josey, 2003; Rupolo et al., 2003b) without much consideration of alternative or concurrent mechanisms such as internal variability. However, one open issue is whether the EMT is to be considered ‘‘unique’’ or whether it is connected to the free internal variability of the EM. Indeed, already in 1961 Wust, in analysing historical data of deep and bottom water in the EM, claimed that ‘‘it seems probable that some smaller influences come from the Aegean Sea by occasional overflow through the channels between Crete and Rhodes. But because of the small number of observations, the conditions of this overflow cannot yet be sufficiently examined ’’. Forty years later Theocharis et al. (2002), basing their analysis on the MEDATLAS database, have documented great variability in the characteristics of Mediterranean water masses during the last century. In particular, there have been at least two distinct events characterized by an increase in both salinity and temperature in the Ionian, Cretan and Levantine sub-basins, one in the early 1970s and the other from the mid to late 1990s (Josey, 2003). The second episode is clearly related to the EMT and its evolution that has been monitored up to the current time. As to the first event, less data are available, and we can only conjecture on its causes and development mechanisms. A suggestive hypothesis is that internal ocean processes might prove to be crucial in determining the basic state of the circulation, onto which the atmosphere exerts its own variable forcing, and that both are important in determining what equilibrium will be reached. From this point of view, air–sea interaction would remain the principal driving force of the Mediterranean circulation, but the occurrence of extreme events would depend on the contemporary presence of the appropriate oceanic conditions. In the WM there is evidence that newly formed deep water has exhibited strong variability since 1959 (Lacombe et al., 1981, 1985). Some studies have observed a positive trend of temperature and salinity of DW (Bethoux et al., 1990; Leaman and Schott, 1991), possibly due to a man-induced reduction in the freshwater supply caused by agricultural activities (Rohling and Bryden, 1992). The contemporary observed changes in DW formation and in its characteristics in the Ionian and Levantine sub-basins (Roether et al., 1996) have provided further evidence for the existence of a trend in heat and freshwater budgets over the entire Mediterranean. The question now is to understand if the observed variability is anthropogenic (e.g. the construction of the Aswan High Dam on the Nile River, Boscolo and Bryden, 2001), or if it is at least partially due to the natural variability of the Mediterranean thermohaline
316 Mediterranean Climate Variability
circulation, of which the EMT may be an example. In this case the budget of salt in the WM is modulated by processes in the Levantine sub-basin. Then the variability of the deep water in the Gulf of Lions may be modulated by the transport of salt trough the Sicily Channel. The Gibraltar Strait provides then the ‘‘choke’’ point needed to understand the interaction between the Mediterranean and Atlantic. To the west, Mediterranean Water is injected at intermediate levels into the North Atlantic Ocean. We have showed that the MOW, and in particular its salinity, plays an important role in the stability of North Atlantic circulation. Moreover, observations show that, starting from the last decade this water is also a source of warming at mid-depths in the North Atlantic, with important implications for global climate change (Potter and Lozier, 2004). Local phenomena, like hydraulic control, tidal flow, mixing rates and pressure differences, contribute to determining the variability of mass and tracer transports between the Mediterranean and the Atlantic. Further studies are needed on the interaction of the thermohaline circulation and the local phenomena within the Gibraltar Strait, especially because salt and temperature anomalies originate from this interaction. If high latitude water continues to freshen (Curry et al., 2003) more investigations are also needed on the MOW impact on the deep convection in the North Atlantic. This issue should be of high priority for climate modelling. Nevertheless the representation of MOW in climate models suffers from many deficiencies. These are mainly related to the difficulty of correctly representing the overflows in large-scale ocean models. In the case of the MOW, the solution to the problem of correctly describing the exchange between the Mediterranean Sea and the North Atlantic in a changing climate situation will constitute a milestone in the description of the coupling between the two basins and their behaviour as a unique system.
References Arhan, M. (1987). On the large scale dynamics of the mediterranean outflow. Deep Sea Res., 34, 1187–1208. Arhan, M., Colin de Verdiere, A., & Memery, L. (1994). The eastern boundary of the subtropical North Atlantic. J. Phys. Oceanogr., 24, 1295. Armi, L., & Zenk, W. (1984). Large lenses of highly saline Mediterranean water. J. Phys. Oceanogr., 14, 1560. Artale, V., Calmanti, S., & Sutera, A. (2002a). Thermohaline circulation sensitivity to intermediate-level anomalies. Tellus A, 54, 159. Artale, V., Iudicone, D., Santoleri, R., Rupolo, V., Marullo, S., & D’Ortenzio, F. (2002b). Role of surface fluxes in ocean general circulation models using satellite
The Atlantic and Mediterranean Sea as Connected Systems
317
sea surface temperature: validation of and sensitivity to the forcing frequency of the Mediterranean thermohaline circulation. J. Geophys. Res., 107(C8), 29. Astraldi, M., Gasparini, G. P., Vetrano, A., & Vignudelli, S. (2002). Hydrographic characteristics and interannual variability of water masses in the central Mediterranean: a sensitivity test for long-term changes in the Mediterranean Sea. Deep-Sea Res. I, 49, 661. Ashkenazy, Y., & Stone, P. H. (2005). Box modeling of the Eastern Mediterranean Sea. J. Phys. Oceanogr., to appear. Beranger, K., Mortier, L., Gasparini, G. P., Gervasio, L., Astraldi, M., & Crepon, M. (2004). The dynamics of the Sicily Strait: a comprehensive study from observations and models. Deep-Sea Research Part II – Topical Studies in Oceanography, 51(4–5), 411. Be´thoux, J., Gentili, B., Morin, P., Nicolas, E., Pierre, C., & 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. Be´thoux, J. P., Gentili, B., Raunet, J., & Tailliez, D. (1990). Warming trend in the Western Mediterranean deep water. Nature, 347, 660. Blanke, B., & Raynaud, S. (1997). Kinematics of the Pacific equatorial undercurrent: an Eulerian and Lagrangian approach from GCM results. J. Phys. Oceanogr., 27, 1038. Blanke, B., Arhan, M., Madec, G., & Roche, S. (1999). Warm water parts in the equatorial Atlantic as diagnosed with a general circulation model. J. Phys. Oceanogr., 29, 2753. Blanke, B., Speich, S., Madec, G., & Do¨o¨s, K. (2001). A global diagnostics of interocean mass transfer. J. Phys. Oceanogr., 31, 1623. Blumberg, A. F., & Mellor, G. L. (1987). A description of a three-dimensional coastal ocean circulation model, Vol. 4: Three-dimensional coastal ocean model, coastal estuarine sciences, N. Heaps (Ed), Amer. Geophys. Union, 1–16. Boscolo, R., & Bryden, H. (2001). Causes of long-term changes in the Aegean Sea deep water. Oceanologica Acta, 24, 519. Bower, A. S., Armi, L., & Ambar, I. (1997). Lagrangian observation of meddy formation during a Mediterranean undercurrent seeding experiment. J. Phys. Oceanogr., 27, 2545. Bower, A. S., Le Cann, B., Rossby, T., Zenk, W., Gould, J., Speer, K., Richardson, P. L., Prater, M. D., & Zhang, H. M. (2002). Directly measured mid depth circulation in the Northeastern North Atlantic. Nature, 419, 403–407. Boyle, E. A., & Keigwin, L. (1987). North Atlantic thermohaline circulation during the past 20000 years linked to high-latitude surface-temperatures. Nature, 330(6143), 35. Brandt, P., Rubino, A., Sein, D. V., Baschek, B., Izquierdo, A., & Backhaus, J. O. (2004). Sea level variations in the Western Mediterranean studied by a numerical tidal model of the Strait of Gibraltar. J. Phys. Oceanogr., 34, 433. Brasseur, P., J. M. Brankart, J. M., Schoenauen, R., & Beckers, J. M. (1996). Seasonal temperature and salinity fields in the Mediterranean Sea: climatological analyses of an historical data set. 43. Deep Sea Res., 43, 159–192. Bray, N. A., Ochoa, J., & Kinder, T. H. (1995). The role of the interface in exchange through the Strait of Gibraltar. J. Geophys. Res., 100, 10755. Castellari, S., Pinardi, N., & Leaman, K. (2000). Simulation of water mass formation processes in the Mediterranean Sea: influence of the time frequency of the atmospheric forcing. J. Geophys. Res., 105(C10), 24157.
318 Mediterranean Climate Variability
Chan, W. L., & Motoi, T. (2003). Effects of stopping the Mediterranean outflow on the southern polar region. Polar Meteorology and Glaciology, 17, 25. Curry, R., Dickson, B., & Yashayaev, I. (2003). A change in the freshwater balance of the Atlantic Ocean over the past four decades. Nature, 426(6968), 826. Delworth, T., Manabe, S., & Stouffer, R. (1993). Interdecadal variations of the thermohaline circulation in a coupled ocean-atmosphere model. J. of Climate, 6(11), 1993–2011. Demirov, E., & Pinardi, N. (2002). Simulation of the Mediterranean Sea circulation from 1979 to 1993. Part 1. The interannual variability. J. of Marine Systems, 33–34, 23. Do¨o¨s, K. (1995). Inter-ocean exchange of water masses. J. Geophys. Res., 100(C7), 13499. Dokken, T. M., & Jansen, E. (1999). Rapid changes in the mechanism of ocean convection during the last glacial period. Nature, 401, 458. D’Ortenzio, F., Marullo, S., & Santoleri, R. (2000). Validation of AVHRR path finder SST’s over the Mediterranean Sea. Geophys. Res. Lett., 27.2, 241–244. Gacic, M., Marullo, S., Santoleri, R., & Bergamasco, A. (1997). Analysis of the seasonal and interannual variability of the sea surface temperature field in the Adriatic Sea from AVHRR data. J. Geophys. Res., C102, 22937–22946. Ganachaud, A., & Wunsch, C. (2003). Large-scale ocean heat and freshwater transport during the world ocean circulation experiment. J. of Climate, 16, 696. Ganopolski, A., & Rahmstorf, S. (2001). Rapid changes of glacial climate simulated in a coupled climate model. Nature, 409, 153. Gerdes, R., Ko¨berle, C., Beckmann, A., Herrmann, P., & Willebrand, J. (1999). Mechanisms for spreading of the Mediterranean water in coarse-resolution numerical models. J. Phys. Oceanogr., 29, 1682. Griffies, S. M., & Tziperman, E. (1995). A linear thermohaline oscillator driven by stochastic atmospheric forcing. J. Climate, 8(10), 2440. Haynes, R., & Barton, E. D. (1990). A pole ward flow along the Atlantic coast of the Iberian Peninsula. J. Geophys. Res., 95, 11425. Hecht, M., Holland, W., Artale, V., & Pinardi, N. (Eds) (1997). North Atlantic model sensitivity to Mediterranean waters. Assessing climate change, Results from the model evaluation consortium for climate assessment, (Wendy Howe and Ann Henderson-Sellers, Gordon and Breach Science, Publishers, 169). Hogg, N. G., & Stommel, H. M. (1990). How currents in the upper thermocline could advects meddies deeper down. Deep-Sea Res., 37, 613. Hopkins, T. S. (1999). The thermohaline forcing of the Gibraltar exchange. J. of Marine System, 20, 1. Hurrell, J. W. (1995). Decadal trends in the North Atlantic Oscillation: regional temperatures and precipitations. Science, 269, 676. Iorga, M. C., & Lozier, S. M. (1999). Signatures of the Mediterranean outflow from a North Atlantic climatology 2. Diagnostic velocity field. J. Geophys. Res., 104(C11), 26011. Iudicone, D., Buongiorno Nardelli, B., Santoleri, R., & Marullo, S. (2003). Distribution and mixing of intermediate water masses in the Channel of Sicily (Mediterranean Sea), J. Geophys. Res.-Oceans, 108(C9). Josey, A. S. (2003). Changes in the heat and freshwater forcing of the Eastern mediterranean and their influence on deep water formation. J. Geophys. Res., 108, doi: 10.1029/2003JC001778.
The Atlantic and Mediterranean Sea as Connected Systems
319
Kinder, T., & Bryden, H. (1990). Aspiration of deep waters through straits. The physical oceanography of sea straits. In: L. J. Pratt, (Ed), (Kluwer, Boston, 295). Kinder, T., & Parrilla, G. (1987). Yes, some of the Mediterranean outflow does come from great depth. J. Geophys. Res., 92, 2901. Klein, B., Roether, W., Manca, B. B., Bregant, D., Beitzel, V., Kovacevich, V., & Luchetta, A. (1999). The large deep water transient in the Eastern Mediterranean. Deep Sea Res.I, 46, 371. Krauss, W., & Ka¨se, R. H. (1984). Mean circulation and eddy kinetic energy in the Eastern North Atlantic. J. Geophys. Res., 89(C3), 3407. La Violette, P. E. (1990). The Western Mediterranean circulation experiment (WMCE): introduction. J. Geophys. Res., 95, 1511. La Violette, P. E. (Ed) (1994). Seasonal and interannual variability of the Western Mediterranean sea: Washington, Am. Geophys. Union, Coastal and Estuarine Studies, 46. Lacombe, H., & Richez, C. (1982). The regime of the Strait of Gibraltar. Hydrodynamics of semi-enclosed seas. In: J. C. J. Nichol, (Ed), (Elsevier, Amsterdam, 13). Lacombe, H., Gascard, J. C., Gonella, J., & Bethoux, J. P. (1981). Response of the mediterranean to the water and energy fluxes across its surface, on seasonal and interannual scales. Oceanologica Acta, 4(2), 247. Lacombe, H., Tchernia, P., & Gamberoni, L. (1985). Variable bottom water in the Western Mediterranean basin. Progress in Oceanography, 14, 319. Lascaratos, A., Roether, W., Nittis, K., & Klein, B. (1999). Recent changes in the Eastern Mediterranean deep waters: a review. Progress in Oceanography, 44, 5. Leaman, K., & Schott, F. (1991). Observations with moored acoustic doppler current profilers in the convection regime in the Gulf of Lions. J. Phys. Oceanogr., 21(4), 558. Lenderink, G., & Haarsma, R. J. (1994). Variability and multiple equilibria of the thermohaline circulation associated with deep-water formation. J. Phys. Oceanogr., 24, 1480. Levitus, S., Antonov, J. I., Boyer, T. P., & Stephens, C. (2000). Warming of the world ocean. Science, 287, 2225. Lozier, M. S., Owens, W. B., & Curry, R. G. (1995). The climatology of the North Atlantic. Progress in Oceanography, 36, 1. Madec, G., Delecluse, P., Imbard, M., & Le´vy, C. (1998). OPA 8.1 ocean general circulation model reference manual. Note du Poˆle de mode´lisation, Institut Pierre-Simon Laplace, n. 11, pp. 91. Malanotte-Rizzoli, P., Manca, B., Ribera d’Alcala, M., Theocharis, A., Bergamasco, A., Bregant, D., Budillon, G., Civitarese, G., Georgopoulos, D., Michelato, A., Sansone, E., Scarazzato, P., & Souvermezoglou, E. E. (1997). A synthesis of the Ionian Sea hydrography, circulation and water mass pathways during POEM Phase I. Progress in Oceanography, 39(3), 153. Malanotte-Rizzoli, P., Manca, B., Ribera d’Alcala`, M., Theocharis, A., Brenner, S., Budillon, G., & Ozsoy, E. (1999). The Eastern Mediterranean in the 80s and in the 90s: the big transition in the intermediate and deep circulations. Dyn. Atmos. Oceans., 29, 365. Malanotte-Rizzoli, P., Manca, B., Marullo, S., Ribera d’Alacala, M., Roether, W., Theocharis, A., Bergamasco, A., Budillon, G., Sansone, E., Civitarese, G., Conversano, F., Gertman, I., Hernt, B., Kress, N., Kioroglu, S., Kontoyannis, H., Nittis, K., Klein, B., Lascaratos, A., Latif, M. A., Ozsoy, E., Robinson, A.,
320 Mediterranean Climate Variability
Santoleri, R., Viezzoli, D., & Kovacevic, V. (2003). The LIWEX Group: The Levantine Intermediate Water Experiment (LIWEX) group: Levantine basin, a laboratory for multiple water mass formation processes. J. Geophys. Res., 108(C9), doi:10.1029/ 2003GB002031. Marullo, S., Santoleri, R., Malanotte-Rizzoli, P., & Bergamasco, A. (1999). The sea surface temperature field in the eastern mediterranean from advanced very high resolution radiometer (AVHRR) data. Part II: interannual variability. J. of Mar. Syst., 20, 83–112. Mariotti, A., Struglia, M. V., Zeng, N. K., & Lau, M. (2002). The hydrological cycle in the Mediterranean region and implications for the water budget of the Mediterranean Sea. J. of Climate, 15(13), 1674. Maze`, J. P., Arhan, M., & Mercier, H. H. (1997). Volume budget of the Eastern boundary layer off the Iberian Peninsula. Deep-Sea Res., 44, 1543. McCartney, M. S., & Mauritzen, C. (2001). On the origin of the warm inflow in the Nordic Sea. Progress in Oceanography, 51, 125. McDowell, S. E., & Rossby, H. T. (1978). Mediterranean water: an intensive mesoscale eddy off the Bahamas. Science, 202, 1085–1087. McManus, J. F., Francois, R., Gherardi, J.-M., Keigwin, L. D., & Brown-Leger, S. (2004). Collapse and rapid resumption of Atlantic Meridional circulation linked to deglacial climate changes. Nature, 428, 834. MEDOC Group (1970). Observation of formation of deep water in the Mediterranean Sea in 1969. Nature, 227, 1037. Mertens, C., & Schott, F. (1998). Interannual variability of deep-water formation in the northwestern Mediterranean. J. Phys. Oceanogr., 1410. Millot, C. (1999). Circulation in the Western Mediterranean Sea. J. of Mar. Syst., 20, 423. Millot, C., Fuda, J.-L., Candela, J., & Tber, Y. (2006). Warming and salting of the mediterranean outflow due to shifts in dense water formation zones. Deep sea Res., in press. Myers, P., & Haines, K. (2000). Seasonal and interannual variability in a model of the Mediterranean under derived flux forcing. J. Phys. Oceanogr., 30, 1069–1082. New, A. L., Barnard, S., Herrmann, P., & Molines, J. M. (2001). On the origin and pathway of the saline inflow to the Nordic Seas: insights from models. Progr. in Oceanogr., 48, 255. Nof, D. (1982). On the movement of deep mesoscale eddies in the North Atlantic. J. Mar. Res., 40, 57. Nof, D. (1991). Lenses generated by intermittent currents. Deep-Sea Res., 38(3), 325. Pacanowski, R., Dixon, K., & Rosati, A. (1991). The G.F.D.L modular ocean model users guide version 1, GFDL Ocean Group, Technical Report No. 2, NOAA/ Geophysical Fluid Dynamics Laboratory, Princeton, NJ. Paillet, J., & Mercier, H. (1997). An inverse model of the Eastern North Atlantic general circulation and thermocline ventilation. Deep-Sea Res., 44, 1293. Pisacane, G., Artale, V., Calmanti, S., & Rupolo, V. (2006). Decadal oscillations in the mediterranean Sea. a result of the overturning circulation variability in the eastern basin? Climate Research, submitted. Potter, R. A., & Lozier, M. S. (2004). On the warming and salinification of the Mediterranean outflow waters in the North Atlantic. Geophys. Res. Lett., 31(1), L01202, doi: 10.1029/2003GL018161.
The Atlantic and Mediterranean Sea as Connected Systems
321
Rahmstorf, S. (1996). On the freshwater forcing and transport of the Atlantic thermohaline circulation. Clim. Dyn., 12, 799. Rahmstorf, S. (1998). Influence of Mediterranean outflow on climate. Eos, 79, 281–282. Reid, J. (1979). On the contribution of the Mediterranean Sea outflow to the NorwegianGreenland Sea. Deep Sea Res., 26, 1199. Richardson, P. L., & Mooney, K. (1975). Mediterranean outflow – simple advection– diffusion model. J. of Phys. Oceanogr., 5(3), 476. Richardson, P. L., Walsh, D., Armi, L., Schroeder, M., & Price, J. F. (1989). Tracking three meddies with SOFAR floats. J. Phys. Oceanogr., 19, 371. Richardson, P. L., Bower, A. S., & Zenk, W. (2000). A census of Meddies tracked by floats. Progress in Oceanography, 45(2), 209–250. Rios, A. F., Perez, F. F., & Fraga, F. (1992). Water masses in the upper and middle North Atlantic Ocean east of the azores. Deep-Sea Res., 39, 645. Rixen, M., Beckers, J.-M., Levitus, S., Antonov, J., Boyer, T., Maillard, C., Fichaut, M., Balopoulos, E., Iona, S., Dooley, H., Garcia, M.-J., Manca, B., Giorgetti, A., Manzella, G., Mikhailov, N., Pinardi, N., Zavatarelli, M., & the Medar Consortium (2005). The Western Mediterranean deep water: a proxy for climate change. Geophys. Res. Lett., 32 (L12608), doi: 10.1029/2005GL022702. Robinson, A., Leslie, W., Theocharis, A., & Lascaratos, A. (2001). Mediterranean Sea circulation, vol.3: Encyclopaedia of ocean sciences, Academic Press., San Diego, CA, p. 1689. Roether, W., & Fuchs, G. (1988). Water mass transport and ventilation in the northeast Atlantic derived from tracer data. Philosophical Transactions of the Royal Society of London, 325, 63. Roether, W., Manca, B, Klein, B., Bregant, D., Georgeopulos, D., Beitzel, V., Kovacevic, V., & Lucetta, A. (1996). Recent changes in Eastern Mediterranean deep waters. Science, 271, 333. Rohling, E., & Bryden, H. (1992). Man-induced salinity and temperature increases in the Western Mediterranean deep water. J. Geophys. Res., 97, 11191. Rupolo, V., Marullo, S., & Iudicone, D. (2003a). Eastern Mediterranean transient studied with Lagrangian diagnostics applied to a mediterranean OGCM forced by satellite SST and ECMWF wind stress for the years 1988–1993. J. of Geophys. Res., 108(C9), art. no. 8121. Rupolo, V., Babiano, A., Artale, V., & Iudicone, D. (2003b). Sensitivity of the Mediterranean circulation to horizontal space-time dependent tracer diffusivity field in a OGCM. II Nuovo Cimento, 26(C4) . Samuel, S., Haines, K., Josey, D., & Myers, P. (1999). Response of the Mediterranean Sea thermohaline circulation to observed changes in the winter wind stress in the period 1980–1993. J. Geophys. Res., 104(C4), 7771. Sannino, G., Bargagli, A., & Artale, V. (2002a). Numerical modelling of the exchange through the Strait of Gibraltar. J. Geophys. Res., 107(C8), 3094, doi:10.1029/ 2001JC000929. Sannino, G., Bargagli, A., & Artale, V. (2002b). Numerical study of the hydraulics of the mean flow through the Strait of Gibraltar, In: The 2nd Meeting on the physical oceanography of sea straits, Villefranche, 15th–19th April 2002 (http://www.soc. soton.ac.uk/JRD/PROC/STRAIT/).
322 Mediterranean Climate Variability
Sannino, G., Bargagli, A., & Artale, V. (2004). Numerical modelling of the semidiurnal tidal exchange through the Strait of Gibraltar. J. Geophys. Res., 109, C05011, doi:10.1029/2003JC002057. Sarnthein, M., Winn, K., Jung, S. J. A., Duplessy, J. C., Labeyrie, L., Erlenkeuser, H., & Ganssen, G. (1994). Changes in East Atlantic deepwater circulation over the last 30,000 years – eight time slice reconstructions. Paleoceanography, 9, 209. Seidov, D., & Haupt, B. J. (1997). Simulated ocean circulation and sediment transport in the North Atlantic during the last glacial maximum and today. Palaeoceanogr., 12, 281. Spall, M. A. (1994). Mechanism for low-frequency variability and salt flux in the Mediterranean salt tongue. J. Geophys. Res., 99, 10121. Spall, M. A. (1999). A simple model of the large-scale circulation of Mediterranean water and Labrador Sea water. Deep-Sea Res., 46, 181. Sparrow, M., Boebel, O., Zervakis, V., Zenk, W., Cantos-Figuerola, A., & Gould, W. J. (2002). Two circulation regimes of the Mediterranean outflow revealed by Lagrangian measurements. J. Phys. Oceanogr., 32(5), 1322–1330. Stanev, E. V. (1992). Numerical experiment on the spreading of Mediterranean water in the North Atlantic. Deep-Sea Res., 39, 1747. Stephens, J. C., & Marshall, D. P. (1999). Dynamics of the Mediterranean salinity tongue. J. Phys. Oceanogr., 29, 1425. Stommel, H. (1961). Thermohaline convection with two stable regimes of flow. Tellus, 13, 224–230. Stommel, H. M., Bryden, H., & Manglesdorf, P. (1973). Does some of the Mediterranean outflow came from great depth? Pure Appl. Geophy., 105, 874–889. Stratford, K., & Haines, K. (2002). Modelling changes in the Mediterranean thermohaline circulation 1987–1995. J. Mar. Syst., 33–34, 51. Theocharis, A., Lascaratos, A., & Sofianos, S. (2002). Variability of sea water properties in the Ionian, Cretan and Levantine seas during the last century, CIESM, Tracking long-term hydrological change in the Mediterranean Sea. CIESM Workshop Series, no. 16, 134 pages, 71, Monaco (www.ciesm.org/publications/Monaco02.pdf). Theocharis, A., Nittis, K., Kontoyiannis, H., Papageorgiou, E., & Balopoulos, E. (1999). Climatic changes in the Aegean Sea influence the Eastern Mediterranean thermohaline circulation (1986–1997). Geophys. Res. Lett., 26, 1617. Thorpe, R. B., & Bigg, G. R. (2000). Modelling the sensitivity of Mediterranean outflow to anthropogenically forced climate change. Climate Dynamics, 16(5), 355. Treguier, A. M., Talandier, C., & Theetten, S. (2003). Modelling Mediterranean water in the North-East Atlantic, Project supported by MERCATOR (2000–2002), Rapport interne LPO 02–14. Tsimplis, M., & Baker, T. (2000). Sea level drop in the Mediterranean Sea: an indicator of deep water salinity and temperature changes. Geophys. Res. Lett., 27, 1731. Tziperman, E. (1987). The Mediterranean outflow as an example of deep buoyancy driven flow. J. Geophys. Res., 92(C13), 14510. Tziperman, E. (2000). Proximity of the present day thermohaline circulation to an instability threshold. J. Phys Oceangr., 30, 90. Van Aken, H. M. (2000). The hydrography of the mid-latitude Northeast Atlantic Ocean II: the intermediate water masses. Deep-Sea Res., 47, 897–924. Vargas-Yanez, M., Ramirez, T., Cortez, D., de Puelles, M. L. F., Lavin, A., LopezJurado, J. L., Gonzales-Pola, C., Vidal, I., & Sebastian, M. (2002). Variability of the Mediterranean water around the Spanish coast: project RADIALES. In F. Briand
The Atlantic and Mediterranean Sea as Connected Systems
323
(Ed), Tracking long term hydrological change in the Mediterranean Sea. CIESM Workshop Series, Vol. 16, pp. 25–28. Whitehead, J. A. (1985). A laboratory study of Gyres and uplift near the strait of Gibraltar. J. Geophys. Res., 90(C10), 7045. Willebrand, J., Barnier, B., Boning, C., Dieterich, C., Killworth, P. D., Le Provost, C., Jia, Y. L., Molines, J. M., & New, A. L. (2001). Circulation characteristics in three eddy-permitting models of the North Atlantic. Progress in Oceanography, 48(2–6), 123. Wu, P., & Haines, K. (1996). Modeling the dispersal of LIW and its role in Mediterranean deep water formation. J. Geophys. Res., 101(C3), 6591. Wu¨st, G. (1961). On the vertical circulation of the Mediterranean Sea. J. Geophys. Res., 66, 3261. Yin, F., & Sarachik, E. S. (1995). Interdecadal thermohaline oscillations in a sector ocean general circulation model: advective and convective processes. J. Phys. Oceanogr., 25, 2465.
This page intentionally left blank
Chapter 6
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment P. Lionello,1 J. Bhend,2 A. Buzzi,3 P.M. Della-Marta,2 S.O. Krichak,4 A. Jansa`,5 P. Maheras,6 A. Sanna,7 I.F. Trigo8 and R. Trigo9 1
University of Lecce, Italy (
[email protected]) University of Bern, Switzerland (
[email protected],
[email protected]) 3 ISAC-CNR, Italy (
[email protected]) 4 Tel Aviv University, Israel (
[email protected]) 5 INM Spain (
[email protected]) 6 University of Thessalonoki, Greece (
[email protected]) 7 ARPA, Piemonte, Italy (
[email protected]) 8 Instituto de Meteorologia/Centro de Geofı´sica, Universidade de Lisboa, Portugal (
[email protected]) 9 CGUL at University of Lisbon and Universidade Luso´fona Portugal (
[email protected]) 2
6.1. Introduction Cyclones represent the most important manifestation of the mid-latitude highfrequency variability, and play a fundamental role in the atmospheric large-scale horizontal (and vertical) mixing and in modulating the air–sea interaction. Cyclonic circulations, due to their frequency, duration and intensity, play an important role in the weather and climate over the entire Mediterranean region (Radinovic, 1987). A large spectrum of environmental variables and phenomena are associated with cyclones in the Mediterranean region. Wind, pressure, temperature, cloudiness, precipitation, thunderstorms, floods, waves, storm surges, landslides, avalanches, air quality and even fog and visibility in the Mediterranean are influenced by the formation and passage of cyclonic disturbances.
326 Mediterranean Climate Variability
Cyclones are important when information more complete than that provided by the average geopotential (or sea level pressure) field is required and when aspects of the probability distribution characterizing the statistics of the atmospheric circulation besides its average values are investigated. Consequently, statistical analysis of cyclones is important especially for the ‘‘tails’’ of probability distributions of variables such as precipitation, winds, waves, storm surges that is the part characterizing extremes values. The Mediterranean area, although located to the south of the main Atlantic storm track that more directly affects western and northern Europe, is quite frequently subjected to sudden events of extreme and adverse weather, often having high social and economic impacts. In fact, though many phenomena associated with cyclones are beneficial from the agricultural, hydrological and economical point of view, some of them are damaging and occasionally disastrous (MEDEX, Jansa` et al., 2001a). The morphology of the territory, with small and steep river basins and highly populated, industrialized and tourist areas, makes the Mediterranean particularly sensitive to the impact of heavy rain and consequent flooding. A report (based on 10 years of data) prepared by the Munich Reinsurance Company collects 166 cases of heavy rainfall and floods and 104 cases of strong wind and storms producing serious damages. The total number of deaths is over 1,900 and the quantified economic losses are over 6,000 MEuro. These figures are certainly underestimates. For Spain alone, and only in four years (1996–1999), the Programme of Natural Hazards of the Spanish Directorate of Civil Defence reported 155 deaths by heavy rain and flood events and 28 deaths by storms and strong winds (Jansa` et al., 2001a) in Greece according to the data published by the Hellenic Agricultural Insurance Organization (ELGA) only for the year 2002 the economic losses due to heavy precipitation, floods, hail and extreme winds were over 180 MEuro. Single disastrous events have been recorded, such as the storm of 4 November 1966 (De Zolt et al., 2006) which hit central and north eastern Italy, causing more than 50 deaths and widespread, huge damages in the eastern Alps, Florence and Venice (where the damage produced by the surge is estimated to be equivalent to 400 KEuros present-day). Over the sea, significant wave height as large as 10 or 11 m are also reported in extraordinary storms, like the 10–11 November 2001 storm in the Western Mediterranean, that produced the destruction of beaches and coastal flooding in northern Mallorca (Gomez et al., 2002) and the big storm at the end of December 1979, that seriously damaged the port of Oran (Algeria). More examples can be found in the MEDEX list of selected cases (Jansa` et al., 2001a). A large portion of such severe weather-related events are associated with cyclones in the Mediterranean (http://medex.inm.uib.es). Since the capability of climate models to reproduce intense cyclones and extreme weather events is limited, it is necessary to determine the links
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
327
between their probability distribution and large-scale patterns or/and indices. In fact, as climate models are more accurate in reproducing the large-scale structure of atmospheric circulation than the statistics of cyclones and extreme events, the prediction of the behaviour of large-scale patterns in future climate scenario could be a robust tool for predicting changes in the intensity and characteristics of cyclones and extreme events. It is important to note that the link between the intensity of cyclones and hazardous extreme events is not simple and different characteristics can be involved as different impacts are considered. The intensity of circulation (winds), precipitation (with consequent floods) and of the cyclone itself (measured as the minimum value of the sea level pressure or the strength of the overall associated circulation) are not necessarily related in a simple (linear) way. An example is the disastrous flood which affected central and Northern Italy during November 1966, characterized by very intense precipitation and high winds, whose central minimum pressure was not remarkably low (De Zolt et al., 2005). Finally, Mediterranean cyclones have also an influence on areas outside the Mediterranean region. Radinovic (1987) suggests that cyclones from the Mediterranean region influence the weather and climate further east in central Europe, in countries such as Hungary, Romania, Ukraine and Russia, and in Asian areas, like Syria, Iraq, Iran, Afghanistan or northern India. At the same time, besides cyclones entering the Mediterranean region from the middle latitude storm track, there is evidence of a significant role played by tropical cyclones, which can produce atmospheric circulation patterns advecting moisture into the Mediterranean region and, occasionally, move into it during a later stage of their life cycle after having experienced a transition to extratropical systems (Pinto et al., 2001; Krichak et al., 2004; Turato et al., 2004). The introductory Section 6.1 of this chapter consists of Subsection 6.1.1 describing the evolution of research on cyclones in the Mediterranean region and their role on climate. Section 6.2 describes the dynamics responsible for the formation and evolution of Mediterranean cyclones (Subsection 6.2.1), the datasets available for their analysis (Subsection 6.2.2), the methods for the cyclone identification and the evaluation of the intensity of the cyclonic activity (Subsection 6.2.3). Section 6.3 describes the climatology of cyclones, their characteristic spatial scales, seasonality, area and genesis mechanisms (Subsection 6.3.1); the relation between cyclonic activity and large-scale climate patterns (Subsection 6.3.2); and the observed trends (Subsection 6.3.3). Section 6.4 describes the effects of cyclones on the Mediterranean environment; it is divided into five subsections, describing their role on precipitation, winds, storm surge, ocean waves and landslides. For each subsection, a description of the mechanism explaining the effects of cyclones on the specific phenomenon
328 Mediterranean Climate Variability
is presented, and climate trends are discussed. The concluding Section 6.5 summarizes the available knowledge on cyclones and on their effects in the Mediterranean region. The outlook Section 6.6 discusses the main open research issues and subjects of ongoing and future research.
6.1.1. ‘‘Historical’’ Notes Pioneering studies that include climatology of cyclones and cyclogenesis in the Mediterranean (in the frame of hemispheric studies) are those by Pettersen (1956) and Klein (1957). They are large-scale studies and show that the western Mediterranean is a distinct and very active area of cyclogenesis with frequent presence of cyclones in winter, in the northern Hemisphere. These studies are based on hand-made analyses and subjective detection of the cyclones. Subsequent studies, using the same techniques, were able to investigate smaller scales, even including mesoscale features (Radinovic and Lalic, 1959; Radinovic, 1987; Genoves and Jansa`, 1989). The annual total number of cyclones detected in the mesoscale studies (Radinovic, 1987 or Genoves and Jansa`, 1989) was, as expected, much larger than the number of cyclones found in larger scale analyses. Since 1990 (Alpert et al., 1990), most of the studies on climatology of the Mediterranean cyclones are based on objective analyses and objective techniques aimed at detecting and tracking the cyclones, but there are also studies based on mixed databases, subjective and objective, like Campins et al. (2000). Several studies have focused on the characteristics of the cyclones in the Mediterranean area, on their dynamics, locations, frequency and temporal variability of cyclogenesis (Buzzi and Tosi, 1989; Alpert et al., 1990a,b; Trigo et al., 1999; Campins et al., 2000; Maheras et al., 2001, 2002, Lionello et al., 2002). A different category of studies have analysed the link between cyclones and environment. Most studies have been focused on precipitation (Trigo et al., 2000; Jansa` et al., 2001b; Kahana et al., 2002; Maheras et al., 2002, 2004), but also other aspects such as storm surges, (Trigo and Davies, 2002; Lionello, 2005), and wind waves (Lionello et al., 2002) have been considered. The impact of extreme weather conditions on landslides occurrence has been attempted over different areas of the Mediterranean. Recent techniques tend to favour the use of satellite and airborne imagery to assess changes in geomorphology. However, the use of historical data obtained from in situ analysis plus local newspapers and interviews is a more reliable tool to establish precise links between the timing of intense rainfall events and concurrent observed landslide episodes (e.g. Zeˆzere et al., 1999; Petrucci and Polemio, 2003; Trigo et al., 2005).
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
329
6.2. Mediterranean Cyclones: Data, Methods and Dynamics This section contains a description of the dynamics responsible for formation and evolution of cyclones in the Mediterranean region, and of data and methodologies used for analysing their climatology.
6.2.1. Dynamics of Cyclones in the Mediterranean Region The cyclones in the Mediterranean region represent a well-distinct element of the global climate. A Mediterranean storm track structure has been put into evidence in different studies and it has been shown that the regular cyclone tracks over the Mediterranean are linked to the baroclinic waveguides (Wallace et al., 1988) and to high rate of alternation between cyclones and anticyclones (Pettersen’s, 1956) in the Mediterranean region (Alpert, 1989). The presence of a separate branch of the Northern Hemisphere storm track crossing the Mediterranean region, with areas of more frequent cyclogenesis in the western Mediterranean and of cyclolysis in the central and eastern Mediterranean, though less intense than the storm track in the Atlantic and Pacific, has been confirmed in recent analysis of the Northern hemisphere (Hoskins and Hodges, 2002). At the same time, there are studies showing that the Mediterranean region is among those presenting the highest concentration of cyclogenesis in the world (Pettersen, 1956; see also Radinovic, 1987, for a general view) during the northern hemisphere winter. Some of them are so intense that they are classified as ‘‘meteorological bombs’’ (Conte, 1986; Homar et al., 2002). According to the conceptual model re-proposed by Hoskins et al. (1985), cyclogenesis occurs when and where a high-level PV (Potential Vorticity) positive anomaly overlaps a low-level potential temperature or PV positive anomaly or a frontal zone. The formation of low-level shallow depressions by orography and thermal contrasts is very frequent in the Mediterranean, owing to the complex topography of the region. Therefore hundreds of cyclonic disturbances can be subjectively and objectively detected in the Mediterranean every year (Campins et al., 2000; Picornell et al., 2001). Actually many of them are shallow depressions which cannot be considered deep cyclones, but, they can play a role in the initiation of actual deep cyclogenesis events. There are three kinds of evolution that can be identified in the Mediterranean: (a) The low-level disturbance does not evolve and remains shallow, weak or moderate and nearly stationary if the upper level PV anomaly is absent or too far away to interact with it (Genoves and Jansa`, 1991). In this case real or deep cyclogenesis does not develop. (b) An upper level PV anomaly will create cyclogenesis when arriving over a frontal zone, just under the maximum PV advection at
330 Mediterranean Climate Variability
high levels, with or without the presence of a depression at low level. Some Mediterranean cyclogenesis could be, at least partially of this type, comprising those cyclones generated at the quasi-permanent Mediterranean border front (Alpert and Ziv, 1989) or at some internal fronts. (c) The low-level disturbance rapidly deepens and cyclogenesis occurs when the upper level perturbation moves close enough to interact with it. The importance of orography and latent heat release in the Mediterranean region has been investigated in several studies. The orography of the region changes quantitatively and qualitatively the baroclinic instability process, usually favouring or ‘‘focusing’’ the cyclogenesis (Speranza et al., 1985). The high frequency of orographically induced low-level disturbances may partially explain the high frequency of real cyclogenesis in the Mediterranean (Genoves and Jansa`, 1991; Jansa` et al., 1994). Latent heat release usually sustains and intensifies most of the cyclogenesis processes. In the Mediterranean region, this effect seems to be quite important in the Eastern Mediterranean, when a Sharav cyclone arrives there from the desert and intensifies over the sea (Alpert and Ziv, 1989). Some cases in the western Mediterranean also have the same evolution (Homar et al., 2002). Enhanced baroclinic instability in saturated air (Fantini, 1995) influenced by latent heat release in an environment convectively stable can be another process contributing to cyclogenesis. In general, the role of latent heat release and diabatic processes is a key issue in the Mediterranean region, though of secondary importance when intense orographic cyclogenesis, both Alpine (Buzzi and Tibaldi, 1978; Dell’Osso and Radinovic, 1984; Speranza et al., 1985; Tibaldi et al., 1990; Stein and Alpert, 1993; Alpert et al., 1995; Buzzi, 1997) or non-Alpine (Garcia-Moya et al., 1989), takes place. Such complicate dynamics and the potential for many different mechanisms favouring cyclogenesis imply that extremely diversified classes of cyclones are present in the Mediterranean region. A tentative list, based partially on the mechanisms producing cyclogenesis and partially on the geographical characteristics, would include lee cyclones, thermal lows, small-scale hurricane-like cyclones, Atlantic systems, African cyclones and Middle East lows. Lee cyclones are triggered by the passage of a major synoptic low-pressure system north of the region, so that their generation is expected to be very sensitive to the location of the storm track above Europe. Lee cyclones develop south of the mountain ridges representing the northern boundary of the Mediterranean region. The Gulf of Genoa is the region of most frequent intense cyclogenesis in the Mediterranean, but lee cyclones are also generated in the Adriatic Sea (Flocas and Karacostas, 1994; Ivanc¸an Picek, 1996), in the Cyprus and Aegean Sea areas (Reiter, 1975; Alpert et al., 1990), in the Black Sea,
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
331
and in several areas of the western Mediterranean sea (Jansa`, 1986). The passage of the same synoptic system can be responsible for successive and distinct cyclogenesis in the Gulf of Genoa, Aegean and Black Sea (Trigo et al., 1999). Correspondingly, the spatial distribution of the frequency of cyclogenesis presents relative maxima in the Cyprus and Aegean region, in the Adriatic, in the Palos-Algerian sea, in the Catalonian-Balearic sea, in the Gulf of Lion and of Genoa (Jansa`, 1986; Trigo et al., 1999). Thermal lows are more frequent in spring and summer and their genesis and lysis are modulated by the daily cycle of temperature. Their occurrence is therefore dependent on the amplitude of this cycle and on the land–sea temperature contrast. Many of them remain shallow depressions confined to the lower troposphere. Modeling studies suggest that they are also generated over sea in autumn and winter, when the land–sea temperature gradient is reversed. Small-scale, hurricane-like cyclones, sometimes called Mediterranean Lows, have been detected over sea. They are a special class of Mediterranean cyclones in which the main source of energy is the great amount of latent heat released in large convective cloud clusters, as in tropical cyclones (Rasmussen and Zick, 1987). They are likely to depend critically on the air–sea temperature difference and on the content of moisture in the atmosphere. Atlantic systems mostly enter in the Mediterranean region from the west and northwest, and cross the Mediterranean during their attenuation phase. Even when their central pressure minimum does not pass directly above the Mediterranean Sea, they affect the Mediterranean weather as they can cause secondary lee cyclones. Northern Africa is the source of many cyclones arriving from the south which often form or intensify south of the Atlas Mountains as lee cyclones. Their formation is more likely to occur in spring and summer when static stability is low. The classification of cyclones in the south-eastern Mediterranean region includes several types of cyclones, with different seasonality and origin, such as the Cyprus Lows, Syrian Lows and Red Sea Troughs. The Cyprus Lows are mostly orographically generated or strengthened. The Syrian low (Kahana et al., 2002) is a particular type of the Cyprus low system so intense that penetrates into some areas of Syria. The Red Sea trough system develops during spring and autumn due to topographic effects in the Red Sea area. Usually it is a warm and shallow trough with a very dry southeasterly flow, but torrential rains occur when an upper cold air trough penetrates southwards above it, due to the extreme instability in this situation (Krichak et al., 1997a,b; Krichak and Alpert, 1998).
332 Mediterranean Climate Variability
6.2.2. Available Sets of Data As mentioned before, pioneering studies on the climatology of cyclones in the Mediterranean were based on hand-made subjective analyses. Due to their own nature, the subjective analysis datasets are in general not available except in the institutions where the analyses were prepared, and often not recoverable for subsequent studies. The subjectivity has the advantage that a hand-made analysis can be even more careful and detailed than objective analyses in some areas, due to appropriate conceptual models and complementary information, etc. At the same time, the subjectivity has the disadvantage of being not homogeneous, neither in space nor in time, because some areas and situations can be systematically poorly or incorrectly analysed due to the lack of specific skill and experience of the authors. In general, the quality of objective analyses has improved to the point where their advantages outweigh those of the subjective analyses. The operational objective analyses are presently the main source of data for performing a climatological analysis of cyclones, although the frequent changes in the procedure (due to changes in the forecasting model, in its resolution, in the data available for its initialization and in the data assimilation method) makes the series quite temporally heterogeneous. Generally, only relatively short samples of operational analyses are homogeneous and can be used for climatic studies. In order to compensate for these inadequacies, homogeneous reanalysis datasets, like ERA15, NCEP or ERA40, have been produced, though with resolution which is coarser than most of the advanced operational analyses (Gibson et al., 1996; Kalnay et al., 1996; Simmons and Gibson, 2000). These homogeneous reanalysis data sets (ERA15, NCEP and ERA40) are the basis for all climatological studies in general and for cyclones in particular. Without them, it would be difficult or even impossible to obtain reliable fields of geopotential or SLP needed for systematic analysis of trends and variability. As a practical compromise, attempting to exploit both advantages of the high resolution of recent operational analyses and of the homogenous multidecadal re-analyses, the MEDEX database contains lists of cyclones identified both in operational analyses (HIRLAM/INM, over the Western Mediterranean, and ECMWF over the whole Mediterranean) and reanalyses (ERA40). Another dataset has been created to investigate the changes and variability in SLP from 1850 as part of the EMULATE project (European and North Atlantic daily to multi-decadal climate variability project). Using Reduced Space Optimal Interpolation (RSOI), daily mean sea level pressure fields with 5 grid spacing have been computed based on homogenized daily SLP station series, ship observations and previously reconstructed SLP fields (Ansell et al., 2005). The dataset consists of daily mean sea level data on a 5 5 grid, and it has a lower
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
333
spatial and temporal resolution than the reanalysis datasets. Though, obviously, it cannot detect mesoscale systems or accurately locate the areas of cyclogenesis, this dataset allows the identification and tracking of synoptic-scale cyclones in the Mediterranean region from 1850 to present. Extreme cyclones can also be identified by their effects which can be measured or recorded. Instrumental data on surges, waves, floods, and winds can be used for reconstruction of time series of past cyclones. A peculiar example is provided by the records of storm surges in Venice, from past chronicle and archives. The highest surges have been reported with a precision which is sufficient to reconstruct the frequency of past floods since the 8th century AD (Camuffo, 1993). The time series shows a succession of periods of recurrent floods, which were particularly intense in the first half of the 16th and of the 18th century, separated by more quiet periods, and a continuous positive trend during the second half of the 20th century. This last part of the time series is very accurate since the tide gauge records have begun in 1872. This example represents the limits and potential of such local reconstructions. On one side, they are a tool for obtaining very long time series in historical times. On the other side, in general it is very difficult to associate past variability to large-scale regimes and make the distinction between large-scale and local processes (the soil subsidence in this case) for which additional information is needed. Also the homogeneity of the time series along its whole extent can be argued, as its reconstruction often involves subjective criteria which need to account for the temporal changes in the level of vulnerability of society, in human activities, resources and technological capability. Finally, it is difficult to attribute the cause of trends without supplementary information. Though floods of Venice occurred many times during its history, the last 50 years represent an unprecedented period of frequent and intense events. However, this is largely explained by the local loss of relative sea level (a combination of ground subsidence and sea level rise) not associated to a trend of storminess (Lionello, 2005).
6.2.3. Methodology In recent years, two distinct approaches have been used to study the storm activity over the North Atlantic and Europe: storm track algorithms and analysis of synoptic variability. Usually, storm track algorithms apply sophisticated methods that can detect the regions of storm development (cyclogenesis) and decay (cyclolysis) as well as the specific paths of each individual storm (Murray and Simmonds, 1991; Serreze et al., 1997; Trigo et al., 1999, 2002; Lionello et al., 2002). Analysis of synoptic variability is a simpler approach, which corresponds to the identification of the synoptic variability using a bandpass filter that retains mainly variability on the 2–8 day period of SLP or 850
334 Mediterranean Climate Variability
or 500 hPa geopotential (Buzzi and Tosi, 1989; Hoskins and Hodges, 2002). This second approach has been widely applied to quantify the synoptic activity associated with a high and low NAO (North Atlantic Oscillation) index (e.g. Rogers, 1997; Ulbrich and Christoph, 1999; Trigo et al., 2002b; Krichak and Alpert, 2005a,b). However, the first technique has also been used to show the areas of significant difference in storm activity between winters with high and low NAO index (Serreze et al., 1997). Figure 107, A–D shows the 1,000 hPa geopotental height standard deviation and associated cyclone trajectories. The analysis is applied to the band-pass filtered fields with cut-off periods at 1 and 7 days. Two monthly periods are considered: January 1966 and 1983, which are characterized with a low and high NAO index, respectively. Both approaches show the existence of a major cyclone variability mode, which is strongly associated with NAO (North Atlantic Oscillation). A climatology of cyclones implies a definition of cyclone and a method of detecting (and describing) it. In both these steps there is a wide margin of arbitrariness, which produces significant differences in the results. When considering subjective analyses and a subjective detection method it was usual to retain a disturbance as a cyclone when it was a minimum of pressure
A
B
C
D
Figure 107: Cyclone trajectories (top panels (A) and (B)) and 1,000 hPA standard deviation (bottom panels (C) and (D)) for low (Jan. 1966, left) and high (Jan 1983, right) monthly NAO index values. Analysis is applied to band-pass filtered data (1–7 day window). Only trajectories of cyclones deeper than 25 m and with duration longer than one day are shown.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
335
surrounded by a closed isobar. Of course, the results depend on the spacing between isobars and of the scale of the map, so that the more detailed the map, the higher the number of cyclones identified. These situations and problems are not avoided by using objective methods. There are different conceptual ways to define a cyclone. An option is to define a cyclone as a relative maximum of vorticity (relative or geostrophic vorticity). Another definition states that a cyclone is a relative minimum of sea level pressure (or geopotential height). The choice of the definition will produce significantly different number of cyclones, and will affect also other details, like their location. Not all the minima of sea level pressure can be retained as cyclones, because many of them are too weak or too close together, so that restrictions have to be imposed (like a threshold for the minimum pressure gradient, a minimum distance between centres, or other constraints), which introduce arbitrariness in the definition of the cyclone and therefore in the results. Even more critical is the effect of the resolution, since low-resolution analyses will permit the detection of relatively large-scale cyclones, but will miss the smaller mesoscale disturbances, while, on the contrary, high-resolution analyses will permit the detection of many small mesoscale disturbances, but could miss the description of the relatively largescale disturbances. In fact, the domain of a large-scale cyclone, defined as the area of positive (geostrophic) vorticity, in a region like the Mediterranean characterized by a complex orography and land–sea distribution, is often broken in multiple fragments, without continuity, due to the existence of many high-low small-scale disturbances, with negative–positive vorticity. The maximum detail in the detection and description of small-scale cyclonic disturbances, is achieved by using high-resolution objective analyses directly, but for a good description of larger scale cyclones, lower resolution analyses have to be used or, alternatively, the original fields of the high-resolution analyses have to be spatially smoothed. The results from high-resolution unfiltered fields and from low-resolution or smoothed fields are dramatically different, even with the same definition of cyclones and the same restrictions to the definition. Figure 108 (Gil et al., 2002) is based on a 3-year (Jun 1998 to May 2001) sample of operational analyses from the ECMWF, consisting of four analyses per day at a T319 resolution, which permits 0.5 latitude–longitude gridded maps. From non-smoothed analyses (right-hand side in the figure), 2,248 cyclones are detected in the Eastern Mediterranean (purple frame) and 2,910 in the Western Basin (blue frame). From smoothed fields, using a Cressman filter with a 200 km radius of influence (left side in the figure), these numbers are reduced to 353 cyclones in the East and 437 in the West. Not only the number of cyclones are totally different, but also some details in the distribution: the very important relative maximum south of the Pyrenees obtained from the original fields disappears when the smoothed fields are used.
Figure 108: Number of cyclones detected in a three-year (Jun 1998 to May 2001) sample of operational analyses from the ECMWF at T319 resolution. Bottom panel shows the results of analysing non-smoothed analyses: 2,248 cyclones are detected in the Eastern Mediterranean (purple frame) and 2,910 in the Western Basin (blue frame). Top panel refers to smoothed fields (a Cressman filter of 200 km of radius of influence has been used): 353 cyclones in the East and 437 in the West (from Gil et al., 2002).
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
337
6.3. Climatology of Cyclones in the Mediterranean The climatology of cyclones in the Mediterranean region is highly influenced by the almost enclosed Mediterranean Sea, which represents an important source of energy and moisture for cyclone development; and by its complex land topography, which plays a major role in steering and deflecting air flows. Moreover, being located within the transition between the subtropical high-pressure belt and the mid-latitude westerlies, the Mediterranean is also subject to strong interannual variability of cyclone activity and, consequently of its precipitation regime, water resources. In the following section, a climatology of Mediterranean cyclones is presented, which decribes the spatial distribution of cyclogenesis and associated mechanisms. The average cyclone characteristics, including their intensity and spatial and temporal scales, as well as their intra- and inter-annual frequency variability, will be described.
6.3.1. Characteristics, Sub-Areas of Cyclogenesis, Seasonality and Generation Mechanisms Mediterranean cyclones are generally characterized by shorter life-cycles and smaller spatial scales than extra-tropical cyclones developed in the Atlantic, as shown in the analysis of storm-tracks derived from 6-hourly near surface fields at T106 (1.125 1.125 ) resolution, available from ERA-15 (ECMWF ReAnalysis) (Trigo et al., 1999). Over 65% of cyclones are within subsynoptic scales, with radius of the order of 550 km or less, considerably smaller than the 1,000–2,000 km values, typical of Atlantic synoptic systems. If the shortest living cyclones (with duration lower than 12 h) are excluded, the average life of cyclones in the Mediterranean region is about 28 h, compared to 3–3.5 days in the Atlantic. Radius and maximum gradient tend to scale with the minimum pressure. In general, cyclones are deeper and have a larger radius in the western than in the Eastern Mediterranean. A recent evaluation (restricted to the western Mediterranean region, Picornell et al., 2001), based on higher resolution fields (computed by HIRLAM at 0.5 resolution) and including short-lived cyclones, produced even smaller space and timescale values. In this dataset the radius of most cyclones is within the 150–350 km range (the mean value is 255 km) and the most intense cyclones have lifetimes of 18–24 h. Deepening rates are mostly lower than 2 hPa (6h) 1, though values as high as 10 hPa (6h) 1 can be observed. Therefore, also the average deepening rates are smaller than in the Atlantic, although the lower latitude at which Mediterranean cyclones develop should be accounted for. Moreover, many cyclones in the Mediterranean region have
338 Mediterranean Climate Variability
null or even negative deepening rate, meaning that they originated in neighbouring regions and cross the Mediterranean during their attenuation phase. A very recent comparison between a sample of Atlantic and West Mediterranean cyclones, made by using high-resolution smoothed fields (HIRLAM/INM 0.5 latitude–longitude with Cressman filter of 200 km), confirms the differences indicated above. The average winter geostrophic circulation of the Mediterranean cyclones is 4 107 m2 s 1 whereas it is 7 107 m2 s 1 for the Atlantic disturbances (Campins, personal communication). Note that the geostrophic circulation combines the size (area) of the cyclones and their geostrophic vorticity, so that can be considered as a measure of their total magnitude. The geography of the region, namely the high orography around the Mediterranean Sea and the existence of embayments and inland seas, determines the relatively small areas where cyclogenesis tends to occur (Table 7; Fig. 109) and the variegate mesoscale structure of Mediterranean systems (Alpert et al., 1990a,b; Trigo et al., 1999, 2002a; Maheras et al., 2001; Picornell et al., 2001). These structures correspond to the mechanisms discussed in the previous Section 6.2.1. Figure 109 shows that the most active areas include the Gulf of Genoa, Iberia, Southern Italy, Northern Africa, Aegean Sea, Black Sea, Cyprus, Middle East; the respective seasons when these areas are most active are indicated in Table 7. Further differentiation can be found inside such areas. In fact, studies based on automated database methods (Campins et al., 2000), have resolved smaller mesoscale structures and identified in the western Mediterranean 7 types of cyclones, on the basis of shape and intensity of the associated circulation. However, despite the use of different methodologies, selection criteria and data sets, most studies (e.g. Alpert et al., 1990a,b; Trigo et al., 1999;
Table 7: Cyclogenetic regions in the Mediterranean area and respective seasons with significant activity (after Trigo et al., 1999); values represent average cyclone radius (km). Area
Seasonality
Radius (km)
Sahara Gulf of Genoa Southern Italy Cyprus Middle East Aegean Sea Black Sea Iberian Peninsula
Spring, Summer Whole year Winter Spring, Summer Spring, Summer Winter, Spring Whole year Summer
530–590 530–380 520 330–460 320–460 500 380–400 410
A
B
C
Figure 109: Number of cyclogenesis events detected per 2.25 2.25 in January (A), April (B), and August (C) from 1979 to 1996 in ECMWF re-analyses (from Trigo et al., 1999).
340 Mediterranean Climate Variability
Figure 110: Annual cycle of the standard deviation of the 1,000 hpa geopotential field. The grey lines show individual years, the black thick line the average annual cycle. Values are based on the band-pass filtered fields (1–7 day cut-off periods), the horizontal black line the average value (Lionello and Zardini, 2005, personal communication). Campins et al., 2000; Maheras et al., 2001; Picornell et al., 2001) agree on the spatial location of cyclone generation. The overall synoptic activity over the entire basin has a well-defined annual cycle, being more intense in the period from November to March which corresponds to the so-called storm season (Fig. 110). Though the temporal and spatial distributions of the Mediterranean cyclones present a large intermonthly variability (Alpert et al., 1990a), the analysis favours the definition of three main seasons: winter, spring and summer. Autumn appears as a transitional period
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
341
with large interannual variability, whose months could be characterized as late summer or early winter. The mechanisms typical of winter cyclogenesis in the Mediterranean exhibit contrasting characteristics with those most common in spring and summer seasons, with intermediate situations in spring and autumn. The 3D structure of the cyclones gives a first clear indication: in the Mediterranean, only the winter cyclones are mostly deep cyclones (reaching the 300-hPa level), while in summer they are mostly shallow (reaching the 850-hPa level). The Laplacian of the temperature at low levels, giving the thermal character of the cyclone, is always negative (warm character), but double in magnitude in summer than in winter (Campins et al., 2005). The Mediterranean cyclones are in general more similar to the Atlantic cyclones in winter than in summer (Campins, personal communication). In winter there are strong links between synoptic upper-troughs and local orography and/or low-level baroclinicity observed over the northern Mediterranean coast. In spring and summer, inland cyclogenesis becomes more frequent and also more sensitive to diurnal forcing (Maheras et al., 2001; Picornell et al., 2001; Trigo et al., 2002a). Winter cyclogenesis occurs essentially along the northern coast in three major areas characterized by strong baroclinicity: the lee of the Alps, when an uppertrough is influenced by the mountains, and over the Aegean and Black Seas, when an upper-trough moves over the relatively warm water basins (Trigo et al., 2002a). The role of orographic cyclogenesis (Buzzi and Tibaldi, 1978) is not limited to the Alps, being also fundamental in the triggering of lows in the Gulf of Lyons, south of the Pyrenees, and also in Southern Italy, south of the Apennines. Over the south-eastern Mediterranean region, the intensity of cyclogenetic activity is to a large extent controlled by large-scale synoptic systems over Europe, particularly by those characterized by mid- and upper-tropospheric southward air-mass intrusions and tropopause-folding effects (Krichak and Alpert, 2003). These processes are often associated with the formation of threedimensional potential vorticity structures, jet streaks and low-level jets conditions over the region to the south of Alps (Buzzi and Foschini, 2000; Liniger and Davies, 2003). In spring, the strengthening of the meridional temperature gradient along the northern African coast favours the development of Saharan depressions. These tend to occur on the lee side of the Atlas mountains, within a region of very weak static stability. Thermal forcing plays an increased role in the genesis and maintenance of Mediterranean Lows in spring and, particularly, in summer. As a result, the life-cycles of summer cyclones, especially those developed over Northern Africa and the Iberian Peninsula, follow the diurnal temperature fluctuations; maximum intensity tends to be reached by late afternoon, and cyclolysis tends to occur mostly by early morning (Maheras et al., 2001;
342 Mediterranean Climate Variability
Picornell et al., 2001; Trigo et al., 2002a). Also the Middle East trough, which is a semi-permanent feature primarily induced by the Asian monsoon acting on a planetary scale (Rodwell and Hoskins, 1996), exhibits the same kind of diurnal see-saw associated with the local thermal cycles (Trigo et al., 2002a). There are several reported cases of very intense storms in autumn–winter and spring months, with severe associated weather (intense rainfall, surges, flash floods), which have either developed or re-intensified over the Mediterranean Sea (e.g. Lee et al., 1988; Ramis et al., 1994; Lagouvardos et al., 1996; Lagouvardos and Kotroni, 1999; Doswell III et al., 1998; Pytharoulis et al., 1999). A fraction of these very intense events develop a hurricane-like structure, feeding on the latent heat release at the sea surface; their frequency, space–time distribution, and interannual variability have not been fully investigated, yet (Pytharoulis et al., 1999).
6.3.2. The Role of Large-Scale Climate Patterns on the Mediterranean Cyclones The Mediterranean region is only partially affected by the North Atlantic storm track, whose main path crosses the Northern Atlantic towards Northern Europe. Consequently, the main mode of the North Atlantic storm track variability, which describes its north–south shift and intensification over the Atlantic, is only marginally related to the frequency and intensity of the cyclones in the Mediterranean region, though Trigo et al. (2000) have demonstrated that an association exists and that it depends on the structural characteristics of the cyclones. In fact, the analysis of low-frequency SLP variability patterns and the frequency of cyclones in the Mediterranean region shows that there are important patterns than the NAO (Krichak and Alpert, 2005a,b). The link between NAO and the position and strength of the storm track in the central Atlantic implies a link between NAO and the frequency of orographic cyclogenesis which is triggered by the passage of Atlantic cyclones. Instead, the bulk of the variability over Central and Southern Europe and over the Mediterranean region is linked to low-frequency patterns, whose centres of actions are localized over Europe and eastern Atlantic (like the East Atlantic/Western Russia pattern (EAWR), Krichak et al., 2000, 2002; Krichak and Alpert, 2005a,b). It has been demonstrated that, depending on the area of the Mediterranean region, a high level of SLP synoptic-scale variability is associated with the positive phases of the SENA (Southern Europe Northern Atlantic) and, to a minor degree, to the SCAN (SCANdinavian) patterns. Therefore, intensity of the cyclogenetic activity in the eastern Mediterranean region is to a large extent controlled by the large-scale
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
343
synoptic processes over Europe and especially by those characterized by mid- and upper-tropospheric southward air-mass intrusions and tropopause folding effects (Krichak et al., 2004). The processes are often associated with the formation of three-dimensional PV structures (PV streamers), jet streaks and low-level jets conditions over the region to the south of the Alps (Buzzi and Foschini, 2001; Liniger and Davies, 2003). These conditions tend to stimulate development of mesoscale convective complexes and Mediterranean cyclones. Intensity, location, duration and orientation of the systems as well as their interdecadal trends in association with those of the main European teleconnection patterns appear to be important elements of the eastern Mediterranean weather and climate trends. Moreover, these teleconnections are defined on a monthly scale, while submonthly large-scale features, such as the well known and relatively frequent Euro-Atlantic blocking (Tibaldi et al., 1997), can influence the trajectory of storm tracks and their associated precipitation fields (Trigo et al., 2004). Winter blocking episodes lasting 10–20 days are associated with a large positive anomaly of the 500 hPa geopotential height above the North Sea and with a higher/lower than average number of cyclones in the Mediterranean/North Sea. Finally, since such large-scale analyses are generally based on relatively coarse resolution fields where the subsynoptic and mesoscale characteristics of the cyclones in the Mediterranean region are poorly reproduced, important components of their variability might not be well described yet.
6.3.3. Trends A counting of cyclone centres (without any differentiation on intensity), based on the NCEP re-analysis, which covers the period 1958–1997, shows a reduction of the number of cyclones in western Mediterranean and an increase in the East (Maheras et al., 2001). Linear fit to the data leads roughly to a 15% increase/ decrease. Changes are not seasonally homogeneous. If only the rainy period (October–March) is considered, a reduction of the number of cyclones is evident also in the Eastern Mediterranean. Other studies suggest a distinction between the increasing trend of weak cyclones and the decreasing trend of strong cyclones in the Western Mediterranean Sea (Trigo et al., 2000). It follows that the positive trend identified in the Northern Hemisphere storm track for the last decades of the 20th century (Chang and Fu, 2002) is not valid in the Mediterranean region. The negative trend is confirmed by the analysis of longer time series (Della-Marta and Bhend, personal communication). A climatology of cyclone activity has been created using a newly compiled dataset of the North
344 Mediterranean Climate Variability
Atlantic–European region from 1850 to 2003 (introduced in Section 6.2.2). Seasonally averaged statistics of the cyclone dynamics have been computed with an objective locating and tracking system developed by Murray and Simmonds (1991). Common to all cyclone tracking algorithms, subjective decisions had to be made regarding the cyclone locating and tracking algorithm parameters. In order to optimize the parameters, tracking results based on EMULATE data (1948–2003) have been compared to tracking results based on daily NCEP data which were transferred to a 5 5 grid (Bhend, 2005). The EMULATE reconstructions in the Mediterranean is shown to be reliable because the RSOI error statistics (see Kaplan et al., 2000) are invariant over time and space for the entire 153 years of the reconstruction indicating that the number of predictors (e.g. station-based observations and marine data) are dense enough to reliably reconstruct SLP on the given grid. See Ansell et al. (2005) for more details on the SLP fields. Significant findings are a marked decrease in winter (DJF) cyclone density over most of the western Mediterranean and an increase in cyclone system density in the eastern Mediterranean for the period 1950–2003 (Fig. 111, A). These findings agree with the results of Maheras et al. (2001). In the longer period, 1850–2003, most of the Mediterranean shows a decrease in cyclone system density (Fig. 111, B). Analysis of the cyclone density time series in the form of a Hovmo¨ller plot shows that the frequency of cyclones over the western Mediterranean is highly variable and exhibits large interannual as well as decadal variability over the last 153 years (Fig. 112).
6.4. Weather Patterns and Mediterranean Environment As it was briefly mentioned in the introduction, cyclones have a great influence on important environmental variables and, particularly, on the timing and magnitude of their extreme values. In general, although not all the extreme weather events in the Mediterranean are related to cyclones and most of the cyclones do not produce extreme weather, it is plausible to assume that Mediterranean cyclones influence most of the high-impact phenomena. Moreover, the high variability of cyclone frequency and intensity, within the Mediterranean Sea and its immediate environments, results in contrasting weather conditions in the region, ranging from large arid areas (e.g. Thornes, 1998; Trigo et al., 2002a) to the greatest annual precipitations totals in Europe, in the Dinaric Alps (Radinovic, 1987; Trigo et al., 2002a). This section shows the correlation of cyclones with rain, winds, waves, surges and even landslides, describes the mechanisms involved and discusses variability and trends of the related phenomena.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
345
A
B
Figure 111: The linear trend in the winter (DJF) cyclone density (A) 1950–2003 and (B) 1850–2003. Cyclone density is the average number of cyclones per unit area at any one time. The trend units are the number of cyclones times 1,000 deg. lat. 2 where deg. lat. is a standard length of 1/360th the circumference of the Earth (from Della-Marta and Bhend, 2005, personal communications).
6.4.1. Precipitation Large quantities of rain require a feeding current of warm and wet air to replace the water removed by precipitation. When vertical stability is close to a critical threshold, such inflow at low levels can favour or lead to instability. Therefore, the eastward and poleward sectors of cyclones are suitable places for
346 Mediterranean Climate Variability
Figure 112: The variability of cyclone density as a function of time and latitude at the longitude of 10 W. The Hovmo¨ller plot y axis defines the latitude in degrees North, the x axis defines the winter season (DJF) from 1850 (bottom) to 2003 (top). Cyclone density is the average number of cyclones per unit area at any given time. Units are the number of cyclones per deg. lat. 2 where deg. lat. is a standard length of 1/360th the circumference of the Earth (from Della-Marta and Bhend, 2005, personal communications). prolonged and intense precipitation. Moreover, orographic upslope lifting is also very effective for producing ascent of warm humid air and persistent rainfall. In fact, in many places the coastal or inner relief intersects a moisture feeding flow and can force upward motion and orographic rain. The local intensity of the precipitation is very much dependent on the path followed by the cyclone and by the amount of available moisture. The heaviest rain events take place when the cyclone path is in such a position that it produces the local convergence of moist Mediterranean air. In the Western Mediterranean, this feeding flow is southerly for northern Italy and Ticino, south-easterly for France, and easterly for Catalonia, the Balearics and Valencia Murcia (Jansa` et al., 2001b). In the Eastern Mediterranean, it is mostly westerly for the Middle East countries and southerly or south-westerly for Greece. These dynamics explain why heavy rain events are associated with cyclones (only a few events in the eastern Mediterranean and in Northern Italy can be an exception) as humid Mediterranean air is advected against the slopes of the mountain ridges surrounding the basin. In fact, synoptic-scale disturbances have been found responsible for most of the floods both in the Western Mediterranean
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
347
and in the Eastern Mediterranean. Only a minority of local flash floods has been associated with intense small convective cells, whose presence is not detected in the standard meteorological analysis. It has been established that for most of the cases (around 90%) of heavy rain in the Western Mediterranean there was a cyclone in the vicinity (Jansa` et al., 2001b, see Fig. 113), though its intensity could vary from an intense and deep system to quite weak and shallow depressions. In fact, considering the mean value for all cyclones identified in such cases, the average vorticity in a central area of the cyclone of 400 km of radius is 0.8 10 4 s 1. In approximately 80% of the events of heavy rain, the location of the cyclonic centre is such that a role of the cyclone in the heavy rain generation and/or location can easily be inferred. The total number of heavy rain events considered is more than 900 (in 5 years, 1992–96). A heavy rain event is defined here as a day with more than 60 mm/day (lowered to 30 mm/day in Algeria) of precipitation in any point of a ‘‘territorial unit’’ (province, department, region or island). Analogously, in Greece about the 92% of rainfall during the rainy period (October–March) is produced by cyclonic patterns (Maheras and Anagnostopoulou, 2003). During the cold season, precipitation in the southern part of the eastern Mediterranean (EM) region is also mainly associated with cyclonic systems of Mediterranean origin. A study carried out for the Negev Desert identified 4 classes of synoptic disturbances responsible for most of the floods, the two most important denoted as the Syrian Low and the Red Sea Trough (Kahana et al., 2002). Therefore cyclones are the cause of most of intense precipitation in the whole Mediterranean region. The cyclones responsible for precipitation in different areas do not share a common origin and generally a single system affects only part of the Mediterranean region. A study focused on Portugal, Italy and Greece has shown that Atlantic Lows dominate in Portugal, where rainy months are associated with an enhanced number of deep and medium cyclones between Newfoundland and British Isles. Precipitation over Greece is very rarely affected by Atlantic cyclones, but is associated to cyclogenesis inside the Mediterranean region. Italy is influenced from both Atlantic and Mediterranean cyclones, because distant lows sometimes contribute to advection of humidity and some Atlantic cyclones may cross over into the Mediterranean and influence the precipitation over Italy. However, the majority of precipitation sources are Mediterranean cyclones (Pinto et al., 1999). During the cold season, precipitation in the southern part of the eastern Mediterranean (EM) region is mainly associated with cyclonic systems of Mediterranean origin. The cyclones producing intense rain in Israel usually belong to the type of Cyprus Lows or Cyprus Depressions (El-Fandy, 1946; Kallos and Metaxas, 1980; Alpert et al., 1995, 2003; Krichak et al., 2004) and usually start their development in the south-western areas of the Mediterranean Sea, and then migrate to the east.
A
B
C
D
E
F
G
H
Figure 113: Most frequent location (elliptic area) of a cyclone in case of heavy rain in some western Mediterranean regions (isolated dot): From left to right and from up to down the locations here considered are (A) SE France, (B) Corsica, (C) North Italy, (D) Catalonia, (E) Balearics, (F) Sardinia, (G) Valencia and (H) Algeria (from Jansa` et al., 2001b).
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
349
During the migration process, the depressions often weaken, though a very significant strengthening of the cyclones is often observed in the Cyprus area, where the lows regenerate in the lee of the Taurus Mountains of Turkey. This phenomenology corresponds to the key role which cyclones play in the internal redistribution of moisture in the Mediterranean region (Fernandez et al., 2003). The transport of moisture from the Western to the Eastern Mediterranean corresponds to the generation (or intensification) of cyclones in the western Mediterranean and their subsequent eastward motion. Air–sea interaction and significant latent heat flux are likely to play an important role in this process. In Greece, for the period 1958–2000, precipitation has been analysed considering the frequency of cyclones and the probability of precipitation produced by them. For the majority of stations, in wintertime the decreasing trends of wet-day amount and the probability of rainfall are consistent with the observed changes in frequency of the various types of cyclones. During autumn on the one hand, the probability of rainfall increases for a large number of cyclonic circulation types, which, on the another hand, are characterized by a decrease in frequency. These opposite trends partially compensate, so that the autumn overall amount of precipitation shows a positive trend (Maheras et al., 2004). An outstanding problem in Western Mediterranean rainfall is the occurrence of catastrophic torrential rains, which tend to occur in the autumn season along coastlines with heavy orography, and the change of relative frequency of moderate/light vs. intense precipitation events. The analysis of observations shows different trends depending on the intensity of the events (Alpert et al., 2002). The torrential rainfall in Italy exceeding 128 mm/day has increased percentage-wise by a factor of 4 during 1951–1995. In Spain, extreme categories at both tails of the distribution (light: 0–4 mm/day and heavy/torrential: 64 mm/day and above) increased significantly. Very little work has been done to link the probability of these events to large-scale extra tropical circulation patterns (Valero et al., 1997), although some analysis of the frequency and intensity of Mediterranean cyclones (Trigo et al., 2002a) and modeling studies for individual events have been performed (Homar et al., 1999; Romero et al., 1999; Pastor et al., 2001). In the last decades, a tendency for more intense concentration of rainfall seems to have occurred along the Mediterranean coastal areas in Italy and Spain (Brunetti et al., 2001; Goodess and Jones, 2002). Over Italy, results show a negative and significant long-term trend in the number of wet days and a positive one in precipitation intensity, which is significant only in the northern regions (Brunetti et al., 2004). The negative trend in wet days persists since the end of 19th century and is due to the marked decrease in the number of low-intensity precipitation events. An increase in the number of events
October - March
200
o
mm / (3.75 x 2.5 )
480
160
400
120
360
80
320
40
Rainfall
o
440
280 0 58 60 62 64 66 68 70 72 74 76 78 80 82 84 86 88 90 92 94 96
Frequency of Mediterranean Cyclonic Events
350 Mediterranean Climate Variability
Figure 114: Time series, and respective linear trends, of the total amount of precipitation in the Northern Mediterranean Basin (bold curve, left axis), the total occurrence of intense Mediterranean cyclones (light curve, right axis), and of non-intense cyclones (dotted curve, right axis) for the October–March period (from Trigo et al., 2000).
belonging to the highest intensity interval was observed too, but only in northern regions. The decrease of total precipitation during the wet season in the Northern Mediterranean has been associated with the reduction of intense cyclones (Fig. 114) and to the northward shift of the storm track over Europe in the period from the 1979 onwards (Trigo et al., 2000). No significant trends were found in Israel, Greece and Cyprus. A detailed analysis of the precipitation in the Valencia Region (Spain) suggests that land use changes in the coastal region result in surface drying, which in turn implies warmer and drier air masses over the coast and higher condensation level and fewer summer storms. It is moreover suggested that higher sea surface temperature can be the cause for the increased number of torrential rain in autumn and winter and that these two factors can be part of a climate mechanism affecting the whole western Mediterranean (Milla´n et al., 2005a,b). The consequent redistribution of the daily rainfall categories – torrential/heavy against moderate/light intensities – is of paramount interest particularly in the semi-arid subtropical regions for purposes of water management, soil erosion and flash floods impacts. Specific isolated regions exhibit an increase of extreme rainfall in spite of the reduction of the total amount of precipitation.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
351
6.4.2. Strong Winds Many strong winds observed in the Mediterranean belong to the category of local winds, like Mistral, Tramontana, Sirocco, Etesian, Bora, Khamsin or Sharav (see H.M.S.O., 1962, or Reiter, 1975, for a general description), that is, they have repetitive location, behaviours and characteristics. Figure 115 shows the location of the main winds in the Mediterranean region. In general, there is a connection between the Mediterranean strong winds and the Mediterranean cyclones. According to the climatology based on ship observations, the highest frequency of gale storm winds in the Mediterranean occur in the Gulf of Lyons, with large difference with other regions (see, for instance, H.M.S.O., 1962), and they can be identified as winds belonging to the Mistral local wind category. The primary cause for these winds is a cyclone located within the Mediterranean, in or near the Genoa region (together with an anticyclone in France or northwestern Europe) and the high frequency and intensity of the Mistral winds is a consequence of the high frequency and intensity of the Genoa cyclones. Similarly, a low pressure above Italy or west of it produces a Sirocco storms in the Adriatic Sea, where the channelling effect of Apennines and Dinaric Alps strongly intensifies a flow which would otherwise be distributed on a larger front. The interaction of air flow and orography contributes to the Mediterranean local winds, which can be partially seen as downslope flows or due to channelling effects. The Mediterranean local winds are attributable to orographic mesoscale pressure perturbation induced by the flow–mountain interaction. High- and low-pressure poles of the orographic disturbance (and/or the orographic pressure dipole as a whole) create local areas of strong pressure gradient that provide intense local acceleration, leading to the local wind generation (Campins et al., 1995). The onset of a local wind is, therefore, often quite abrupt. Past the narrow accelerating zone, the winds continue blowing and spreading in an inertial way, although density gradients can contribute to a more efficient wind spreading and extension (Jansa`, 1960; Alpert et al., 1982). According to this mechanism, local winds are shallow, only 1.5–2 km deep at most (Jansa`, 1933; Campins et al., 1995; Saaroni et al., 1998) and may remain quite independent from the basic flow above the mountain. On the other hand, the frequent presence of intense cyclones is enough to explain some windstorms blowing within the Mediterranean region and the effect of orographic forcing is not always fundamental. The hot winds blowing from the desert across the Libyan and Egyptian segments of flat coast (Chili, Shimum), the strong easterlies over the Eastern Mediterranean (Saaroni et al., 1998), and the Libeccio storms in the Tyrrhenian Sea, could constitute situations where such interaction is not crucial. Note that the extraordinary storms of
352 Mediterranean Climate Variability
Figure 115: Maps of the lagged correlation with the value in the point indicated by the dot for the 500 hpa synoptic-scale filtered geopotential height. This point is selected as a maximum of explained variance of the precipitation field associated to the EOF describing a moisture transport from western to eastern Mediterranean areas. The maps are meant to describe the evolution of the cyclonic disturbances associated with such transport (From Fernandez et al., 2003).
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
353
December 1979 and November 2001, mentioned in Section 6.0, belong to the category of ‘‘intense cyclones’’. It can be added that the indirect methods of estimating winds (the scatterometers carried by satellite, like in ‘‘Quickscat’’) has given for the sustained winds in one of these storms (November 2001) values as high as 35 m/s. The synergistic combination of both mechanisms, that is the presence of intense cyclone and the generation of local winds, explains the high speed of extreme wind events. Etesian winds blow over the Aegean Sea and belong to a larger circulation system including northerly winds over the whole Eastern Mediterranean. This Northerly circulation is produced by the combination of a high pressure over Balkan Peninsula or over central Europe and low pressure over eastern Mediterranean and Iraq, generally of thermal origin and corresponding to an extension of the summer Tibetan low to the eastern Mediterranean sea (Repapis et al., 1977). The Etesian winds can be intensified by the presence of a trough in the mid or upper troposphere over eastern Mediterranean. The temperature contrast between land and the Aegean sea can also influence (intensify or weaken) the intensity of the Etesian winds (Maheras, 1980). It is also worthy to remark that the leading edge of wind streams can act as an internal shallow front. On the other hand, associated with local wind streamers is the formation of orographically generated cyclonic and anticyclonic PV banners, characterized mainly by shear vorticity which often contributes to the Mediterranean cyclogenesis stimulating in some cases the heavy rainfall events (Ae¨bischer, 1996; Ae¨bischer and Scha¨r, 1996).
6.4.3. Storm Surge Storm surge results from the combined action of atmospheric pressure and wind stress on the sea surface. Atmospheric pressure produces what is called the inverse barometric effect, according to which, in steady conditions a low pressure is associated with a sea level increase. Wind stress pushes horizontally the water column and tends to accumulate it at the closed end of a basin. In steady condition, the slope of the sea surface is proportional to wind stress and to the inverse of the water depth. Therefore, the action of wind stress dominates in shallow water. These dynamics explain the importance of cyclones for storm surges and why the variability of cyclone regimes has an impact on the surge events which, in turn, can be considered an indicator of the cyclones characteristics. The storm surges in the northern Adriatic, and the consequent flooding of Venice, is caused by intense cyclones in the north-western Mediterranean (Trigo and Davies, 2002; Lionello, 2005). The synoptic patterns determining the surge
354 Mediterranean Climate Variability
BORA
MISTRAL TRAMONTANA
MAESTRALE VARDARIS SIROCCO ETESIAN SHARKIYA
LIBECCIO
VANDEVALES
SIROCCO ETESIAN SIROCCO SHARAV
CHILL GHIBLI
KHAMSIN
Figure 116: Location and direction of main winds in the Mediterranean region. Note that there are discrepancies between related terms in the different languages. In the French language, Mistral is the northerly wind descending along the Rhone Valley and blowing in the region of Provence, towards the western Mediterranean and Tramontane the northwesterly wind belonging to the same system and blowing in the Roussillon region. In the Italian language Tramontana is a northerly wind and Maestrale the dominant northwesterly wind blowing over most of the western Mediterranean Sea. Tramontana (or Tramuntana) is also the name of the northerly wind (belonging to the Mistral system) in Catalan and Spanish languages, blowing in the north of Catalonia and north of the Balearics.
in the Gulf of Venice present a low-pressure system with a minimum above central Europe or northern Italy (Figs. 116, 117) which produces a strong Sirocco wind along the Adriatic Sea. Although these synoptic dynamics are well known, the mechanisms responsible for their frequency and intensity have not been completely understood. It appears that periods with extreme surge events are characterized by a general circulation anomaly, represented by a pattern with a negative centre of action above the Eastern Atlantic, according to which cyclones are deviated south-eastward and penetrate into the
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
A
B
C
D
355
E
Figure 117: Evolution of the gph1000 field producing the storm surge in Venice. Panels (A) to (E) show the gph1000 composites (in hPa according to the greyscale bar) with a time lag of 48h, 24h, 0h, þ24h, þ48h with respect to the time of the surge peak. This situation is associated to a main minimum south of the Alps at the time of maximum surge (panel C). The composites are based on events with a peak surge value higher than 70 cm.
Mediterranean Sea from North–West or generate a lee cyclone south of the Alps. This pattern is different from the NAO dipole, whose time behaviour is not correlated with that of the highest surges in the Gulf of Venice. It appears therefore that the storminess associated with the floods of Venice is not related to the NAO (Fig. 118 after Lionello, 2005). The records of floods in Venice show a clear positive trend, the cause of which is mainly the subsidence of the ground level, whose rate was particularly high in the years from the 1950s to the 1970s, rather than variations in the meteorological forcing fields. If subsidence and sea level rise are excluded, the residual
356 Mediterranean Climate Variability
A
B
C
D
E
Figure 118: Same as Fig. 117, but for a situation associated with a main pressure minimum located north of the Alps at the time of maximum surge (panel C).
variability, which can be associated with the meteorological forcing, shows trends which are small and dependent on the intensity of the storm surges. During the second half of the 20th century, there is an indication that the frequency of moderate surges is increasing (Pirazzoli and Tomasin, 1999) while major independent surge events do not show large variation (Trigo and Davies, 2002). During the same period, a weak decreasing trend has been identified in the value of extreme levels in Trieste (Raicich, 2003). This behaviour has been shown to be consistent with the variation of the meteorological forcing in the Northern Adriatic area, that is, with more frequent moderate storms and less frequent intense storms. However, if the effect of regional sea level rise is subtracted from the data, the record of extremes is dominated by a large interdecadal variability, with respect to which an eventual residual trend is of
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
357
Figure 119: Indexes of the storm surge level time (black continuous), sunspots number (grey continuous), NAO (black dashed) during the 1940–2000 period (years on the x axis). NAO and storm surge level data include November and December only. All time series have been smoothed using a 3-year running mean.
minor importance (Fig. 119, Lionello, 2005). The understanding of this variability, and its past and future evolution, appears to be a very interesting scientific and practical issue. It will be important, on one side, to investigate the variability of large-scale patterns associated with it and, on the other side, to find the explanation for its correlation with the periodicity of the sunspot number. This last correlation points to the possible existence of a link between regional climate and external forcing which is not understood yet.
6.4.4. Wind Waves Ocean waves are the consequences of winds, so that intense cyclones are the cause of extreme waves. In this respect storm surges and waves have a common cause. However, waves grow continuously under the action of the wind and their maximum height reflects the average intensity of the wind along the fetch. In other words, waves tend to depend on the integral of the wind stress along their travelling path, while surges depend mostly on its value over the shallow water areas near the coast. Measurements with moored buoys, and models capable of assimilating, analysing and forecasting waves, have demonstrated that high waves (with 5–7 m significant wave height) exist in the Mediterranean, in spite of the relatively short fetches with respect to the Oceanic situations.
358 Mediterranean Climate Variability
As already mentioned, analysed SWH (Significant Wave Height) of 10–11 m were encountered in the case of the extraordinary storm of 10–11 November 2001 (Gomez et al., 2002). It is not simple to reconstruct the wave climatology in the Mediterranean sea because long time-series of instrumental observations are lacking. Buoy observations are mostly available since the 1990s, when national buoy networks were installed. Satellite altimeter data are continuously available only since 1992. Therefore, the analysis of past variability is mostly based on model reconstructions and ship observations. Often, model simulations under-evaluate the SWH, mostly because of lack of accuracy of the forcing wind fields. An example is Fig. 120 which results from a model integration based on the ERA40 reanalysis winds. Figure 120 represents the geographical distribution of maximum SWH values, but presumably does not show the actual height of the more intense events. There are several compilations and atlas of winds and waves derived from ship observations. There is a big difference between the Atlantic and Mediterranean situation, mainly due to different fetch, which in the Atlantic Ocean is generally larger than in the Mediterranean Sea. Therefore, the Atlantic waves are larger (longer period) than those in the Mediterranean, even when the same wind conditions are considered. Due to the low dissipation in long waves travelling out of the storm area (swell), the Atlantic waves remain relatively high for
Figure 120: Variation of the 10-year return value (y axis, cm.) during the 1940–2000 period. The horizontal dashed line shows the value computed on the basis of the whole 1040–2001 period. The grey and black continuous lines show the values computed using events inside a moving 7-year and 21-year long time window, respectively.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
359
a long time, and, therefore, the average wave heights in the Atlantic are much higher than in the Mediterranean, where the swell component plays a minor role so that waves are mostly present only in correspondence to strong winds (windsea). In the Gulf of Lyons, in January, the frequency of waves above 1m of SWH is as large as 70%, and the frequency of waves above 2.5 m of SWH south-east of this zone is 20%. The growth of waves with fetch implies that the maximum frequency of high waves is displaced downstream from the maximum frequency of strong winds so that the highest waves are just downstream of the mistral core. Significant wave heights of 6m or more are reached every year (on average) in this zone of the Mediterranean where their presence is related to the high frequency and intensity of the Genoa cyclones. Figure 121 confirms the presence of these maxima and shows also other features in the various basins. High waves are present over most of the Mediterranean Sea and tend to reach the highest values where strong wind and long fetch are simultaneously present. The largest maxima are located in the western Mediterranean and in the Ionian Sea, under the action of the Mistral, where the shape of the Mediterranean Sea determines the most effective combination of a long fetch and a strong wind. The Island of Crete interrupts the fetch of the Etesian winds and determines two separated maxima: one in the Aegean and another in the Levantine Basin. Sirocco produces the maximum SWH in the Northern Ionian and in the southern Adriatic. A maximum due to the Bora wind is present in the northern Adriatic Sea, and another due to the Vendavel in the Alboran Sea. In summer, wave height is small over the whole basin and a characteristic maximum is present in the Aegean sea caused by the action of the Etesian winds. Figure 122 shows the sea level pressure composites associated with high waves in different areas of the Mediterranean Sea. Cyclones located near Cyprus are responsible for high waves in the Levantine Basin, those in the Gulf of Genoa for high waves in the western Mediterranean, while those above Tyrrenian or central Italy produce high waves in the Adriatic and central Mediterranean Sea. Obviously, waves are the result of the action of past winds, so that also the past evolution of the synoptic situation contributes to the value and location of the maximum SWH. Wind wave extremes, obtained from model simulations, show little significant trends. Figure 123 shows the statistically significant variations of maximum SWH in 50 years, on the basis of linear trends derived from the ERA40 reanalysis. The two main features correspond to a reduction in the Ionian and Alboran Sea, which are consistent with a reduction of cyclones in the western Mediterranean. Increase is limited to a very small region near the coast of France.
360 Mediterranean Climate Variability
Figure 121: Distribution of maximum wave height as resulting from a 40-year model integration carried out using the ERA-40 re-analysis. Contour levels show the maximum SWH, arrows show the mean wave direction corresponding to SWH maximum (from Sanna and Lionello, 2005).
6.4.5. Landslides Rainfall-induced landslides are usually directly associated with the passage of intense storms of Atlantic or Mediterranean origin. In a recent work, the link between the occurrence of Landslide episodes in Portugal and the storm tracks associated with the NAO pattern has been established (Trigo et al., 2005). Naturally, strong cyclones can produce intense rainfall events that are responsible for the rapid growth of pore pressure and for the loss of the apparent cohesion of thin soils, resulting in failure within the soil material or at the contact with the underlying impermeable bedrock. A different type of association between cyclones and landslides can be found for rainfall periods which are less intense but have a long duration. In this case long-lasting rainfall periods (from 30 days to 90 days), are responsible for the activity of deeper slope movements, such as translational slides, rotational slides and complex and composite slope movements. This is the group of landslide events mostly affected by the large-scale atmospheric circulation mode NAO. The western Mediterranean Basin is prone to slope instability due to geological, geomorphological and climatic factors. It is widely accepted that high
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
A
B
C
D
361
E
Figure 122: Synoptic patterns associated with extreme significant wave height in different regions of the Mediterranean Sea: (A) Tyrrhenian, (B) Adriatic, (C) Balearic, (D) Ionian, (E) Levantine basin (from Sanna and Lionello, 2005).
duration/intensity rainfall events (associated with intense cyclone events) are the most important triggering mechanism of landslides worldwide (van Asch et al., 1999). In particular, rainfall-induced landslides have been studied in Portugal (Trigo et al., 2005), Spain (Corominas and Moya, 1999), Italy (Polemio and Petrucci, 2000) and France (Flageollet et al., 1999). Landslide consequences include damages on property, houses and particularly roads, and can be also related with the increasing human pressure related to urban development throughout the countryside (Trigo et al., 2005).
362 Mediterranean Climate Variability
Figure 123: Statistically significant variations of maximum significant wave height in 50 years, based on linear trends evaluated from the results of a 40-year model integration carried out using the ERA-40 re-analysis.
6.5. Conclusions The Mediterranean region is characterized by strong morphological forcing. The steep orography surrounding the basin and the complicated land–sea distribution introduce a rich mesoscale structure. On the one hand, these two factors produce a peculiar phenomenology internal to the basin and, on the other hand, they modulate the interaction of the Mediterranean system with global climate patterns by adding sub-regional features. As far as cyclones are concerned, steep orography and a complex land–sea distribution condition the formation and the evolution of cyclones themselves and the effects they produce on the environment. Mountain ridges and land–sea contrast are responsible for the presence of many areas with orographic cyclogenesis, and thermal lows, respectively. The presence of moist Mediterranean air with potential for strong diabatic processes conditions the development of cyclonic structures, like small-scale, hurricane-like lows. The triggering mechanism of most Mediterranean cyclones is mostly due to features external to the Mediterranean region, that is mid-latitude (primary) baroclinic waves with high-level potential vorticity anomaly, which interacts with regional structures in the Mediterranean region. The complexity of the mechanisms involved is such that many different categories of cyclones can be identified, according to the region of formation, their seasonality and dominant mechanism of formation. Beside the cyclones entering from the Atlantic sector, there are lee cyclones, thermal lows, African cyclones, mesoscale hurricanelike lows, Middle East cyclones. Further differentiations are possible, as several
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
363
distinct regions of cyclogenesis exist for lee cyclones and different types of cyclones can be defined in the Middle East. Most cyclones have a radius smaller than 600 km. Seasonality is different for various categories; in general, the overall synoptic activity is higher from November to March, but there are types of cyclones, like thermal lows, Sahara cyclones and Middle East depressions, whose frequency is larger in summer. Because of this complexity, several large-scale patterns can be associated with cyclones in the Mediterranean area. While the NAO certainly plays an important role (which is dependent on the type of cyclone considered), blocking episodes above central and northern Europe explain a large fraction of variability. Other patterns, like the East Atlantic/Western Russia pattern, exert a significant influence too. Long-term trend analysis (since 1850) shows a reduction of the cyclone activity over most of the region. During the second half of the 20th century this trend is confirmed for the western part, while an increase of cyclone activity has been observed for the eastern part. However, a very large interannual variability is superimposed on these trend. Cyclones are associated with many extreme events: precipitation, winds, waves, landslides and surges. Several hazardous weather events take place every year in the Mediterranean region and are a relevant cause of economic losses. In all these events, the geography at regional-scale plays a fundamental role in determining the effects of cyclones on the environment. Prolonged heavy rain episodes take place when a cyclone forces surface currents of humid and warm Mediterranean air to flow over coastal mountain slopes. High waves are produced when the location of the cyclone ensures a long fetch. Strong winds occur when orography locally intensifies the cyclonic circulation around a lowpressure centre. The storm surge in the Northern Adriatic sea takes place when the south-easterly Sirocco wind is channelled along the Adriatic Sea. In general, this implies that the intensity of the cyclone is not the only factor responsible for its impact, but its position, evolution and track are also extremely important for the impact on the environment. In general, extremes present larger variability than average values and, consequently, it is more difficult to identify significant trends. While most of the Mediterranean region, in winter, experiences a decrease in total precipitation and average SWH (Lionello and Sanna, 2005), extremes do not show spatially and temporally coherent trends over the whole Mediterranean region. Extreme SWH levels have become smaller only in part of the Ionian and in the Alboran Sea, while are increasing in a small area close to the coast of France. The frequency of torrential rainfall has been found to increase (percentually wise) in the second half of the 20th century in some areas of the western Mediterranean (Alpert et al., 2002).
364 Mediterranean Climate Variability
6.6. Outlook This chapter documents the large amount of research which has already been carried out and provides a well-established understanding of many aspects of the Mediterranean cyclones and their effect on the environment. However, in consideration of the importance of these processes, of their potentially damaging effects, and the need for assessing their sensitivity to climate changes, further research is needed (see also Chapter 8). A main issue is whether the present archives of data provide an adequate database for the required analysis. On the centennial timescales it appears that there is an imbalance between regions where phenomena are well documented (e.g. the surge of Venice, Camuffo, 1993), and regions where data are scarce and reconstruction of past events necessarily indirect (e.g. precipitation patterns and extremes on the whole African coast). For the recent decades, where meteorological observations are available worldwide and model reanalysis have been carried out (e.g. NCEP and ERA-40 re-analysis), it has still to be fully investigated whether the subsynoptic and mesoscale characteristics of cyclones in the Mediterranean region are well represented in the available data archives. In this respect, the development of extensive, high-resolution sets of data appears extremely important for addressing unresolved scientific issues. Similar considerations apply for climate change studies. Certainly more work is needed to link modeling of selected events, long-term modeling and observation analysis, with the goal of a coherent long-term climatological perspective. In this respect, interactions between climate and meteorological projects, such as MedCLIVAR (endorsed by the WCRP project CLIVAR) and MEDEX (a project of WMOWWRP), national initiatives and the involvement of regional institutions are certainly important and potentially fruitful. The classifications of cyclones should involve an analysis of the sensitivity of the generation mechanisms to climate variability and change, whose identification might help in predicting future scenarios and the change in the frequency and intensity of some specific cyclone types (e.g. the hurricane-like Mediterranean Lows). The availability of regional high-resolution reanalysis, where cyclone structures are well reproduced, might be crucial for this task. Moreover, it is important to identify the deficiencies in the models that account for the inadequacies in simulating the cyclones and their variability. On this point, the connection between the climate of the cyclones in the Mediterranean region and the low-frequency large-scale variability is not sufficiently understood yet. Intensity, location, duration and orientation of the systems as well as their interdecadal trends should be analysed and put in relation with variations of the main European teleconnection patterns. The importance of regional-scale processes (e.g. latent heat release over the sea) and their variability with respect
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
365
to large-scale forcing (e.g. the meridional shift and/or intensification of the storm track over northern Europe) has not been precisely quantified. The Mediterranean Sea is characterized by a complex coastline structure, with some highly vulnerable areas (for example, the Niles delta and the Gulf of Venice). Though the main danger for these areas is due to the increased coastal erosion and loss of land that would be produced by sea level rise, the change in storminess is also potentially critical, because variations of the frequency and intensity of sea storms could further increase risks and damages. The analysis of changing wind waves and surge regimes requires detailed impact studies, carried out on the basis of sufficiently precise forcing fields, and relies on accurate surface wind field analysis and adequate downscaling techniques. Variations of cyclone regimes affect the distribution of precipitation. It may be suggested that the variations of the precipitation observed during the last decade over the Mediterranean region are associated with relatively small variations in the transport and characteristics of the air masses. Such changes might be small and not always easily detectable; however, their impact on local climate and climate variability is likely to be large. These variations could have serious consequences for rain intensities in many Mediterranean areas. The danger is twofold. There are areas already under stress because of recurrent water shortage during summer, and areas where torrential rains have produced human casualties and large damages to properties. It is important to identify the factors responsible for the increase in rainfall extremes and reduction of total precipitation. A similar understanding is also important for waves, surges and winds. Finally, the links between large-scale patterns and extreme events are not simple, as extreme events are cannot easily associated to extreme values of some large-scale predictors. The characterization of patterns of cyclones, weather extremes and their link to large-scale fields is a topic on which more research is needed.
Acknowledgements The comments by P. Malguzzi have been of great help in improving the content of this chapter. The authors are indebted to C. Zerefos, P. Alpert, J. Luterbacher and E. Xoplaki for important suggestions and to E. Elvini for his help with the graphics.
References Ae¨bischer, U. (1996). Low-level potential vorticity and cyclogenesis to the lee of the Alps. PhD Thesis, Dissertation No 11732, Swiss Federal Institute of Technology (ETH), Zu¨rich.
366 Mediterranean Climate Variability
Ae¨bischer, U., & Scha¨r, C. (1996). Low-level potential vorticity and cyclogenesis to the lee of the Alps. MAP Newsletter, 5, 68–69. Alpert, P. (1989). Baroclinic waveguides and rate of alternation. J. Atmos. Sci., 46, 3505–3507. Alpert, P., & Ziv, B. (1989). The Sharav cyclone, observations and some theoretical considerations. J. Geophys. Res., 94, 18495–18514. Alpert, P., Cohen, A., Neumann, J., & Doron, E. (1982). A model simulation of the summer circulation from the Eastern Mediterranean past lake Kinneret in the Jordan Valley. Mon. Wea. Rev., 110, 994–1006. Alpert, P., Neeman, B. U., & Shay-El, Y. (1990a). Climatological analysis of Mediterranean cyclones using ECMWF data. Tellus, 42A, 65–77. Alpert, P., Neeman, B. U., & Shay-El, Y. (1990b). Intermonthly variability of cyclone tracks in the Mediterranean. J. Climate, 3, 1474–1478. Alpert, P., Stein, U., & Tsidulko, M. (1995). Role of sea fluxes and topography in Eastern Mediterranean cyclogenesis. The Global Atmosphere-Ocean System, 3, 55–79. Alpert, P., Osetinsky, I., Ziv, B., & Shafir, H. (2004). Semi-objective classification for daily synoptic systems: application to the Eastern Mediterranean climate change. Int. J. Climatol., 24, 1001–1011. Alpert, P., Ben-Gai, T., Baharad, A., Benjamini, Y., Yekutieli, D., Colacino, M., Diodato, L., Ramis, C., Homar, V., Romero, R., Michaelides, S., & Manes (2003). The paradoxical increase of Mediterranean extreme daily rainfall in spite of decrease in total values. Geophys. Res. Lett., 29, 31-1–31-4. Ansell, T. J., Jones, P. D., Allan, R. J., Lister, D., Parker, D., Brunet, M., Moberg, A., Jacobeit, J., Brohan, P., Rayner, N., Aguilar, E., Alexandersson, H., Barriendos, M., Brandsma, T., Cox, N. J., Della-Marta, P. M., Drebs, A., Founda, D., Gerstengarbe, F., Hickey, K., Jo’nsson, T., Luterbacher, J., Nordli, O., Oesterle, H., Petrakis, M., Philipp, A., Rodwell, M. J., Saladie, O., Sigro, J., Slonosky, V., Srnec, L., Swail, V., Garcı´ a-Sua´rez, A. M., Tuomenvirta, H., Wang, X., Wanner, H., Werner, P., Wheeler, D., & Xoplaki, E. (2005). Daily mean sea level pressure reconstructions for the European-North Atlantic region for the period 1850–2003. J. Climate, accepted. Bhend, J. (2005). North Atlantic and European Cyclones: their variability and change from 1881 to 2003. Masters Thesis, University of Bern. Brunetti, M., Colacino, M., Maugeri, M., & Nanni, T. (2001). Trends in the daily intensity of precipitation in Italy from 1951 to 1996. Int. J. Climatol., 21, 299–316. Brunetti, M., Maugeri, M., Monti, F., & Nanni, T. (2004). Changes in daily precipitation frequency and distribution in Italy over the last 120 years. J. Geophys. Res., 109, D05102. Buzzi, A. (1997). Finite amplitude and moisture effects in orographic cyclones. INM/ WMO International Symposium on Cyclones and Hazardous Weather in the Mediterranean, Palma de Mallorca, 14–17. Buzzi, A., & Tibaldi, S. (1978). Cyclogenesis on the lee of Alps: a case study. Q-Q.J.R. Meteorol. Soc., 104, 171–287. Buzzi, A., & Tosi, E. (1989). Statistical behaviour of transient eddies near mountains and implications or theories of lee cyclogenesis. J. Atmos. Sci., 46, 1233–1249. Buzzi, A., & Foschini, L. (2000). Mesoscale meteorological features associated with heavy precipitation in the Southern Alpine region. Meteorol. Atmos. Phys., 72, 131–146.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
367
Campins, J., Genove´s, A., Jansa`, A., Guijarro, J. A., & Ramis, C. (2000). A catalogue and a classification of surface cyclones for the Western Mediterranean. Int. J. Climatol., 20, 969–984. Campins, J., Jansa`, A., Benech, B., Koffi, E., & Bessemoulin, P. (1995). PYREX observation and model diagnosis of the Tramontane wind. Meteorol. Atmos. Phys., 56, 209–228. Campins, J., Jansa`, A., & Genove´s, A. (2005). Three dimensional structure of west mediterranean cyclones. Int. J. Climatol., submitted. Camuffo, D. (1993). Analysis of sea surges at Venice from A.D. 782 to 1990. Theor. Appl. Climatol., 47, 1–14. Chang, E. K. M., & Fu, Y. (2002). Interdecadal variation in Northern Hemisphere winter storm track intensity. J. Climate, 15, 642–658. Conte, M. (1986). The Meteorological bomb in the Mediterranean: a synoptic climatology. WMO/TD No. 128, App. 4, 17–31. Corominas, J., & Moya, J. (1999). Reconstructing recent landslide activity in relation to rainfall in the Llobregat River basin, Eastern Pyrenees, Spain. Geomorphology, 30(1–2), 79–93. Dell’Osso, L., & Radinovic, D. (1984). A case study of cyclone development in the lee of the Alps on 18 March 1982. Beitr. Phys. Atmos., 57, 369–379. De Zolt, S., Lionello, P., Malguzzi, A., Nuhu, A., & Tomasin, A. (2005). The 4th November 1966 storm over Italy and its effects on wind waves and storm surge, in preparation. De Zolt, S., Lionello, P., Malguzzi, A., Nuhu, A., & Tomasin, A. (2006). The disastrous storm of 4th November 1966 on Italy, in preparation. Doswell III, C. A., Ramis, C., Romero, R., & Alonso, S. (1998). A diagnostic study of three heavy precipitation episodes in the western Mediterranean region. Wea. Forecasting, 3, 102–124. El-Fandy, M. G. (1946). Barometric lows of Cyprus. Q.J.R. Meteorol. Soc., 72, 291–306. Fantini, M. (1995). Moist Eddy waves in a quasigeostrophic three-dimensional model. J. Atmos. Sci., 52, 2473–2485. Fernandez, J., Saenz, J., & Zorita, E. (2003). Analysis of wintertime atmospheric moisture transport and its variability over the Mediterranean basin in the NCEPReanalyses. Clim. Res., 23, 195–215. Flageollet, J. C., Maquaire, O., Martin, B., & Weber, D. (1999). Landslides and climatic conditions in Barcelonnette and Vars basins (Southern French Alps, France). Geomorphology, 30(1–2), 65–78. Flocas, H. A., & Karacostas, T. S., (1994). Synoptic and dynamic characteristics of cyclogenesis over the Aegean Sea. International Symposium on the Life Cycles of Extratropical Cyclones, Bergen, Norway, 186–191. Garcia-Moya, J. A., Jansa`, A., Diaz-Pabon, R., & Rodriguez, E. (1989). Factor influencing the Algerian sea cyclogenesis. WMO/TD No., 298, 87–94. Genove´s, A., & Jansa`, A. (1989). Statistical approach to mesoscale non-alpine West Mediterranean cyclogenesis. WMO/TP num, 298, 77–85. Genove´s, A., & Jansa`, A. (1991). The use of potential vorticity maps in monitoring shallow and deep cyclogenesis in the Western Mediterranean. WMO/TD num, 420, 55–65. Gibson, R., Ka˚lberg, P., & Uppala, S. (1996). The ECMWF re-analysis (ERA) project. ECMWF Newsl., 73, 7–17.
368 Mediterranean Climate Variability
Gil, V., Genoves, A., Picornell, M. A., & Jansa`, A. (2002). Automated database of cyclones from the ECMWF model: preliminary comparison between West and East Mediterranean basin. Mediterranean Storms, 4th Plinius Conference 2002 (CD-ROM, available at Centro Meteorologico en Illes Balears, Spain, and at University of the Balearic Islands). Gomez, M., Alvarez, E., Carretero, J. C., Perez, B., Rodriguez, I., Serrano, O., Sotillo, M. G. (2002). Oceanographic and atmospheric analysis of the 10–16 November 2001 Storm in the Western Mediterranean. IV Plinius Conference on Mediterranean Storms (Mallorca, October 2002), CD-ROM (Available at the University of the Balearic Islands and at the Territorial Meteorological Centre in the Balearics, Palma). Goodess, C. M., & Jones, P. D. (2002). Links between circulation and changes in the characteristics of Iberian rainfall. Int. J. Climatol., 22, 1593–1615. Homar, V., Ramis, C., Romero, R., Alonso, S., Garcı´ a Moya, J. A., & Alarco´n, M. (1999). A case of convection development over the Western Mediterranean Sea: a study through numerical simulations. Meteorol. Atmos. Phys., 71, 169–188. Homar, V., Ramis, C., & Alonso, S. (2002). A deep cyclone of African origin in the Western Mediterranean: diagnosis and numerical simulation. Ann. Geophys., 20, 93–106. Hoskins, B. J., McIntyre, M. E., & Roberston, A. W. (1985). On the use and significance of isentropic potential vorticity maps. Quart. J. Roy. Met. Soc., 111, 877–946. Hoskins, B. J., & Hodges, K. I. (2002). New perspectives on the Northern Hemisphere winter storm track. J. Atmos. Sci., 59, 1041–1061. H.M.S.O., Meteorological Office (1962). Weather in the Mediterranean, Pub. 391, Vol. 1, General Meteorology, London. Jansa`, J. M. (1933). Contribucion al estudio de la Tramontana en Menorca. Servicio Meteorolo´gico Espan˜ol, Serie A, num, 3, Madrid. Jansa`, J. M. (1960). Choques de presio´n en las irrupciones frı´ as. Rev. Geofı´sica., 18, 35–50. Jansa`, A. (1986). Genoa cyclones and other Western Mediterranean cyclones. WMO/TD No. 128, App. 8, 59–70. Jansa`, A., Radinovic, D., Alpert, P., Genoves, A., Campins, J., & Picornell, M. A. (1994). Mediterranean cyclones: subject of a WMO Project. International Symposium on the Life Cycle of Extratropical Cyclones, Bergen, Vol II, 26–31. Jansa`, A., Alpert, P., Buzzi, A., & Arbogast, P. (2001a). MEDEX, cyclones that produce high impact weather in the Mediterranean, available at http://medex.inm.uib.es. Jansa`, A., Genove´s, A., Picornell, M. A., Campins, J., Riosalido, R., & Carretero, O. (2001b). Western mediterranean cyclones and heavy rain. Part 2: statistical approach. Meteorol. Appl., 8, 43–56. Kahana, R., Ziv, B., Enzel, Y., & Dayan, U. (2002). Synoptic climatology of major floods in the Negev Desert, Israel. Int. J. Climatol., 22, 867–882. Kallos, G., & Metaxas, D. A. (1980). Synoptic processes for the formation of Cyprus lows. Rivista Meteorologia Aeronautica, 2–3, 121–138. Kalnay, E. et al. (1996). The NCEP/NCAR 40-year re-analysis project. Bull. Amer. Meteor. Soc., 77, 437–471. Kaplan, A., Kushnir, Y., & Cane, M. A. (2000). Reduced space optimal interpolation of historical marine sea level pressure: 1854–1992*. Journal of Climate, 13, 2987–3002.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
369
Klein, W. H. (1957). Principal tracks and frequencies of cyclones and anticyclones in the Northern Hemisphere. U.S. Weather Bur., Res. Paper num 40. Krichak, S. O., & Alpert, P. (1998). Role of large scale moist dynamics in November 1–5, 1994. Hazardous mediterranean weather. Journal of Geophysical Research, 103, 19, 453–19, 458. Krichak, S. O., & Alpert, P. (2005a). Decadal trends in the East Atlantic/West Russia pattern and the Mediterranean precipitation. Int. J. Climatol., 25, 183–192. Krichak, S. O., & Alpert, P. (2005b). Signatures of the NAO in the atmospheric circulation during wet winter months over the Mediterranean region. Theor. Appl. Climatol., in press. Krichak, S. O., Alpert, P., & Krishnamurti, T. N. (1997a). Interaction of topography and tropospheric flow – A possible generator for the red sea trough? Meteorology and Atmospheric Physics, 63, 149–158. Krichak, S. O., Alpert, P., & Krishnamurti, T. N. (1997b). Red Sea trough/cyclone development – numerical investigation. Meteorology and Atmospheric Physics, 63, 159–170. Krichak, S. O., Tsidulko, M., & Alpert, P. (2000). Monthly synoptic patterns associated with wet/dry Eastern Mediterranean conditions. Theor. Appl. Climatol., 65, 215–229. Krichak, S. O., Kishcha, P., & Alpert, P. (2002). Decadal trends of main Eurasian oscillations and the Mediterranean precipitation. Theor. Appl. Climatol., 72, 209–220. Krichak, S. O., Alpert, P., & Dayan, M. (2004). Role of atmospheric processes associated with hurricane Olga in December 2001 flash floods in Israel. J. Hydrometeorol., 5, 1259–1270. Lagouvardos, K., & Kotroni, V. (1999). Aspects of the storm of 25/26 March 1998 over Greece. Weather, 54, 154–160. Lagouvardos, K., Kotroni, V., Dobricic, S., Nickovic, S., & Kallos, G. (1996). The storm of October 21–22, 1994, over Greece: observations and model results. J. Geophys. Res., 101, 26,217–6,226. Lee, T. P.-J., Silberberg, S. R., & Bosart, L. F. (1988). A case study of a severe winter storm in the Middle East. Q. J. R. Meteorol. Soc., 114, 61–90. Liniger, M. A., & Davies, H. C. (2003). Substructure of a MAP streamer. Q. J. R. Meteorol. Soc., 129, 633–651. Lionello, P., Dalan, F., & Elvini, E. (2002). Cyclones in the Mediterranean region: the present and the doubled CO2 climate scenarios. Clim. Res., 22, 147–159. Lionello, P., Elvini, E., & Nizzero, A. (2003). Ocean waves and storm surges in the Adriatic Sea: intercomparison between the present and doubled CO2 climate scenarios. Clim. Res., 23, 217–231. Lionello, P. (2005). Extreme surges in the Gulf of Venice. Present and future climate Venice and its lagoon, State of Knowledge. C. Fletcher, and T. Spencer (Eds), (Cambridge University Press, Cambridge UK, pp. 59–65). Lionello, P., & Sanna, A., (2005). Mediterranean wave climate variability and its links with NAO and Indian Monsoon. Clim. Dynamics, in press. Maheras, P., (1980). Le proble`me des Ete´siens. Me´diterrane´e, N40, 57–66. Maheras, P., Flocas, H., Patrikas, I., & Ch., Anagnostopoulou, (2001). A 40 year objective climatology of surface cyclones in the Mediterranean region: spatial and temporal distribution. Int. J. Climatol., 21, 109–130.
370 Mediterranean Climate Variability
Maheras, P., Flocas, H. A., Anagnostopoulou, Ch., & Patrikas, I. (2002). On the vertical structure of composite surface cyclones in the Mediterranean region. Theor. Appl. Climatol., 71, 199–217. Maheras, P., & Anagnostopoulou, Chr. (2003). Circulation types and their influence on the interannual variability and precipitation changes in Greece, Mediterranean climate-variability and trends (Springer Verlag, Berlin, Heidelberg, pp. 215–239). Maheras, P., Tolika, K., Anagnostopoulou Chr., Vafiadis, M., Patricas, I., & Flocas, H. (2004). On the relationships between circulation types and rainfall variability changes in Greece. Int. J. Climatol., 24, 1695–1712. Milla´n, M. M., Estrela, M. J., Sanz, M. J., Mantilla, E., Martı´ n, M., Pastor, F., Salvador, R., Vallejo, R., Alonso, L., Gangoiti, G., Ilardia, J. L., Navazo, M., Albizuri, A., Artı´ n˜ano, B., Ciccioli, P., Kallos, G., Carvalho, R. A., Andre´s, D., Hoff, A., Werhahn, J., Seufert, G., & Versino, B. (2005a). Climatic feedbacks and desertification. The Mediterranean model. J. Climate, 18, 684–701. Milla´n, M. M., Estrela, M. J., & Miro`, J. J. (2005b). Rainfall components: variability and spatial distribution in a Mediterranean area (Valencia Region), J. Climate, 18, 2682–2705. Murray, R. J., & Simmonds, I. (1991). A numerical scheme for tracking cyclones centres from digital data. Part I: development and operation of the scheme. Aust. Meteor. Mag., 39, 155–166. Nicolaedes, K. (2005). Study of the synoptic and dynamic parameters of the cyclones during the cold period over the Cyprus area. PhD Thesis, pp. 278 (in Greek). Pastor, F., Estrela, M. J., Penarrocha, D., & Millan, M. M. (2001). Torrential rains on the Spanish Mediterranean coast: modelling the effects of the sea surface temperature. J. Appl. Meteorol., 40, 1180–1195. Petrucci, O., & Polemio, M. (2003). The use of historical data for the characterization of multiple damaging hydrogeological events. Natural Hazards and Earth System Sciences, 3, 17–30. Pettersen, S. (1956). Weather analysis and forecasting. Mac Graw Hill, New York. Picornell, M. A., Jansa`, A., Genove´s, A., & Campins, J. (2001). Automated database of mesocyclones from the Hirlam(INM)-0.5 analyses in the Western Mediterranean. Int. J. Climatol., 21, 335–354. Pinto, J. G., Ulbrich, U., & Speth, P., (1999). The variability of cyclonic activity in the Mediterranean area in the last 40 years and its impact on precipitation, In: Proceedings of the 1st EGS Plinius Conference, Maratea, Italy, October 1999, pp. 29–40. Pinto, J. G., Klawa, M., Ulbrich, U., Rudari, R., Speth, P., (2001). Extreme precipitation events over northwestern Italy and their relationship with tropical-extratropical interactions over the Atlantic, In: Proceedings of the 3rd EGS Plinius Conference, Baja Sardinia, Italy, October 2001. Pirazzoli & Tomasin, A. (1999). Evoluzione delle cause recenti dell’Aqua Alta. Atti Istituto Veneto Scienze Lettere ed Arti, CLVII, 317–344. Polemio, M., & Petrucci, O. (2000). Rainfall as a landslide triggering factor: an overview of recent international research. In: Landslides in Research, Theory and Practice E. Bromhead, et al.(Eds), Vol. 3, (Thomas Telford, London, pp. 1219–1226). Pytharoulis, I., Craig, G. C., & Ballard, S. P. (1999). Study of the hurricane-like Mediterranean cyclone of January 1995. Phys. Chem. Earth (B), 24, 627–633. Radinovic, D. (1987). Mediterranean cyclones and their influence on the weather and climate, WMO, PSMP Rep. Ser. num 24.
Cyclones in the Mediterranean Region: Climatology and Effects on the Environment
371
Radinovic, D., & Lalic, D. (1959). Cyclonic activity in the West Mediterranean. Federal Hydromet. Institute, Memories No. 7, Belgrade, 57 pp. Raicich, F. (2003). Recent evolution of sea-level extremes at Trieste (Northern Adriatic). Cont. Shelf. Res., 23, 225–235. Ramis, C., Llasat, M. C., Genove´s, A., & Jansa`, A. (1994). The October 1987 floods in Catalonia: synoptic and mesoscale mechanisms. Meteorol. Appl., 1, 337–350. Rasmussen, E., & Zick, C. (1987). A subsynoptic vortex over the Mediterranean with some resemblance to polar lows. Tellus, 39A, 408–425. Reiter, E. (1975). Handbook for forecasters in the Mediterranean. Part I: general description of the meteorological processes. Naval Environmental Research Facility, Monterey, California. Repapis, C., Zerefos, C. S., & Tritakis, B. (1977). On the Etesians over the Aegean. Proc. Acad. Athens, 52, 572–606. Rodwell, M. J., & Hoskins, B. J. (1996). Monsoons and the dynamics of deserts. Q. J. R. Meteorol. Soc., 122, 1385–1404. Rogers, J. C. (1990). Patterns of low-frequency monthly sea level pressure variability (1899–1986) and associated wave cyclone frequencies. J. Climate, 3, 1364–1379. Rogers, J. C. (1997). North Atlantic storm track variability and its association to the North Atlantic oscillation and climate variability of the Northern Europe. J. Climate, 10, 1635–1647. Romero, M. J., Sumner, G., Ramis, C., & Genove´s, A. (1999). A classification of the atmospheric circulation patterns producing significant daily rainfall in the Spanish Mediterranean area. Intern. J. Climatol., 19, 765–785. Saaroni, H., Ziv, B., Bitan, A., & Alpert, P. (1998). Easterly wind storms as an environmental hazard over Israel. Theoretical and Appl. Climatology, 59, 61–77. Sanna A., & Lionello, P. (2005). Characteristics and trends of significant wave height extremes In the Mediterranean Sea, in preparation. Serreze, M. C., Carse, F., Barry, R. G., & Rogers, J. C. (1997). Icelandic low cyclone activity: climatological features, linkages with the NAO, and relationships with recent changes in the Northern Hemisphere circulation. J Climate, 10, 453–464. Simmons, A. J., & Gibson, J. K. (2000). The ERA-40 Project Plan, ERA-40 Project Report Series n.1. Speranza, A., Buzzi, A., Trevisan, A., Malguzzi, P., (1985). A theory of deep cyclogenesis in the lee of the Alps. Part I: modifications of baroclinic instability by localized topography. J. Atmos. Sci., 42, 1521–1535. Speranza, A., Buzzi, A., Trevisan, A., & Malguzzi, P. (1985). A theory of deep cyclogenesis in the lee of the Alps. Part I: modifications of baroclinic instability by localized topography. J. Atmos. Sci., 42, 1521–1535. STARDEX project (2005). STatistical and Regional Dynamical downscaling of EXtermes for European region. Contract N EVK2-CT-2001–00115. Stein, U., & Alpert, P. (1993). Factor separation in numerical simulations. J. Atmos. Sci., 50, 2107–2115. Thornes, J. B. (1998). Mediterranean desertification, p. 2–4. In: P. Mairota, J. B. Thornes and N. Geeson (Eds), Atlas of Mediterranean Environments in Europe (Wiley, Chichester, 205 pp.). Tibaldi, S., Buzzi, A., & Speranza, A. (1990). Orographic cyclogenesis. In: C. W. Newton and E. O. Holopainen (Eds), Extratropical Cyclones – The Eric Palmen Memorial Volume (American Meteorological Society, Boston, 107–127).
372 Mediterranean Climate Variability
Tibaldi, S., D’Andrea, F., Tosi, E., & Roeckner, E. (1997). Climatology of Northern Hemisphere blocking in the ECHAM model. Clim. Dyn., 13, 649–666. Tosi, E., Fantini, M., & Trevisan, A. (1983). Numerical experiments on orographic cyclogenesis: relationship between the development of the lee cyclone and the basic flow characteristics. Mon. Wea. Rev., 111, 799–814. Trigo, I. F., Davies, T. D., & Bigg, G. R. (1999). Objective climatology of cyclones in the Mediterranean region. J. Climate, 12, 1685–1696. Trigo, I. F., Davies, T. D., & Bigg, G. R. (2000). Decline in Mediterranean rainfall caused by weakening of mediterranean cyclones. Geophysical Research Letters, 27, 2913–2916. Trigo, I. F., & Davies, T. D. (2002). Meteorological conditions associated with sea surges in Venice: a 40 year climatology. Int. J. Climatol., 22, 787–803. Trigo, I. F., Bigg, G. R., & Davies, T. D. (2002a). Climatology of cyclogenesis mechanisms in the Mediterranean. Mon. Wea. Rev., 130, 549–649. Trigo, R. M., Osborn, T. J., & Corte-Real, J. M. (2002b). The North Atlantic Oscillation influence on Europe: climate impacts and associated physical mechanisms. Clim. Res., 20, 9–17. Trigo, R. M., Trigo, I. F., DaCamara, C., & Osborn, T. J. (2004). Climate impact of the European winter blocking episodes from the NCEP/NCAR reanalyses. Clim. Dyn., 23, 17–28. Trigo, R. M., Zeˆzere, J. L., Rodrigues, M. L., & Trigo, I. F. (2005). The influence of the North Atlantic Oscillation on rainfall triggering of landslides near Lisbon, Natural Hazards, 36, 331–354, doi: 10.1007/S11069-005-1709-0. Turato, B., Reale, O., & Siccardi, F., (2004). Water Vapor Sources of the October 2000 Piedmont flood. J. Hydrometeorol., 693–712. Ulbrich, U., & Christoph, M. (1999). A shift in the NAO and increasing storm track activity over Europe due to anthropogenic greenhouse gas. Clim. Dyn., 15, 551–559. Valero, F., Luna, M. Y., & Martin, M. L. (1997). An overwiew of a heavy rain event in Southeastern Iberia: the role of large-scale meteorological conditions. Ann. Geophysicae, 15, 494–502. Van Asch, T., Buma, J., & VanBeek, L. (1999). A view on some hydrological triggering systems in landslides. Geomorphology, 30(1–2), 25–32. Wallace, J. M., Lim, G., & Blackmon, M. L. (1988). Relationship between cyclone tracks, anticyclone tracks and baroclinic waveguides. J. Atmos. Sci., 45, 439–462. Zeˆzere, J. L., Ferreira, A. B., & Rodrigues, M. L. (1999). The role of conditioning and triggering factors in the occurrence of landslides: a case study in the area north of Lisbon (Portugal). Geomorphology, 30(1–2), 133–146.
Chapter 7
Regional Atmospheric, Marine Processes and Climate Modelling Laurent Li,1 Alexandra Bozec,2 Samuel Somot,3 Karine Be´ranger,2 Pascale Bouruet-Aubertot,2 Florence Sevault3 and Michel Cre´pon2 1
LMD/IPSL/CNRS, Universite´ P. et M. Curie Paris, France (
[email protected]) LOCEAN/IPSL, Universite´ P. et M. Curie Paris, France (
[email protected],
[email protected],
[email protected],
[email protected]) 3 Me´te´o-France, CNRM/GMGEC/EAC, Toulouse, France (
[email protected],
[email protected]) 2
7.1. Introduction The Mediterranean region is rather unique in respect to its geographical position: north of the largest desert in the world – the Sahara, and south of a large temperate climate region – Europe. It is therefore a transition area between tropical and mid-latitude climates. As a transition area, the Mediterranean region shows important local climate variability and rather large gradients, both in the South–North and East–West directions. The Mediterranean climate is characterized by its strong seasonal contrast. The summer is dry and hot, the winter is humid and mild. The upper panel of Fig. 124 shows the sea-level pressure for the region of the North Atlantic, Europe and Mediterranean for December–January–Feburary as described in the ERA-15 dataset. The remarkable structure of this figure is the Icelandic Low and the Azores High. The main atmospheric center of action affecting the Mediterranean climate is the Azores High, a subtropical anticyclone related to the descending branch of the Hadley cell. The Mediterranean region can thus be related to tropical climate events like El Nin˜o and monsoons (see also Chapter 2). The Mediterranean Sea is an important playground for the North Atlantic Oscillation, a major atmospheric circulation pattern of the Northern
374 Mediterranean Climate Variability
Figure 124: Sea-level pressure (hPa) and 300-hPa zonal wind (m/s) for December–January–February, as depicted in the ECMWF re-analysis dataset from 1979 to 1993.
Regional Atmospheric, Marine Processes and Climate Modelling
375
Hemisphere, characterized by a seesaw between the Icelandic Low and the Azores High. The Mediterranean climate may thus be strongly influenced by processes which can involve the atmosphere only or coupled ocean–atmosphere phenomena in mid and high latitudes (see also Chapters 3 and 6). As depicted in the lower panel of Fig. 124 showing the 300-hPa zonal wind for the whole Eurasian continent and North Africa, the Mediterranean Sea is located in the north flank of the sub-tropical jet stream. The jet stream plays an important role in forming atmospheric teleconnections between the Mediterranean and regions far away. The Mediterranean Sea is a concentration basin with an evaporation rate much larger than the rainfall rate and river runoff (Mariotti et al., 2002; Struglia et al., 2004), leading to increase in salt content. It is also a source of heat to the atmosphere with annual decreases of temperature for water masses. This particular behaviour of the Mediterranean Sea has its roots and consequences in the Gibraltar Strait, where the inflow is fresh (36.6 psu) and warm (maximum of 21 C in August and minimum of 17 C in March), and the outflow is salty (38.25 psu) and cold (13.3 C). The Mediterranean Sea is thus similar to a thermodynamic engine which transforms the inflowing light Atlantic water into dense deep Mediterranean waters through air–sea coupling (see Chapters 4 and 5 for more descriptions). This water transformation process generates thermohaline forcing which drives, in a large proportion, the Mediterranean marine general circulation. Convection can thus be observed in several places of the Mediterranean Sea, particularly, in the Gulf of Lions, Adriatic Sea, Aegean Sea and Levantine basin. The surface circulation in the western Mediterranean can be schematically described as follows. The Atlantic Water (AW) enters the Alboran Sea forming the Alboran gyres (Gascard and Richez, 1985; Heburn and La Violette, 1990), and flows eastward forming the Algerian Current (AC). The AC presents well-marked meanders due to baroclinic instabilities. Then the AC splits into two branches at the Sicily Strait, one entering the Tyrrhenian Sea passing through the Corsica Strait and forming the Northern current, the other entering the eastern Mediterranean (Millot, 1987; Herbaut et al., 1998). The AW entering the eastern Mediterranean divides into two distinct streams (Robinson et al., 1999; Lermusiaux and Robinson, 2001). One flows over the Tunisian shelf, the other forms the Mid Ionian jet. These two currents merge at the level of East of Libya (as seen in Marullo et al., 1999) as a coastal current (Alhammoud et al., 2005; Hamad et al., 2002) flowing eastwards along the Egyptian coast. Then this current flows northwards along the Jordanian–Israel–Lebanon Coast and westwards at the level of Turkey. During its eastward progression, the AW is transformed through convection processes into Western Mediterranean Deep Water (WMDW) in the Gulf of Lion, into Levantine Intermediate Water
376 Mediterranean Climate Variability
(LIW) in the Levantine Basin, into Eastern Mediterranean Deep Water (EMDW) in the Adriatic and Aegean Seas. The LIW flows westwards into the Western Mediterranean at intermediate depth (400 m) through the Sicily Strait and then into the Atlantic Ocean through the Gibraltar Strait, closing the water budget of the Mediterranean Sea. Numerical modelling, both global and regional, is an important tool to understand physical mechanisms controlling climate change and variability at different spatio-temporal scales. It also provides the unique possibility to construct physically based and comprehensive future climate scenarios, the starting point for many socio-economical impact considerations. Sections 7.2, 7.3 and 7.4 will present several studies on the physical mechanisms controlling the Mediterranean climate variation and change. Sections 7.5 and 7.6 will then present the current status of the Mediterranean regional climate modelling and the preliminary results of a regional coupled model. Perspectives will be given in Section 7.7.
7.2. Teleconnection Patterns from the Mediterranean Region The Mediterranean Sea plays an important role in determining the climate of the nearby regions (Millan et al., 2005a,b). It is also believed that the Mediterranean Sea can exert influences on the climate of regions far away. The first mechanism may be through the Mediterranean outflow water (about 1 Sverdrup of warm and salty water) flowing out of the Gibraltar Strait into the North Atlantic. The Mediterranean Sea can thus contribute to the global climate variation by altering the oceanic overturning circulation (see Chapter 5). Teleconnection patterns in the atmosphere can also be initiated from the Mediterranean region. Rowell (2003) reported that a warming of the Mediterranean Sea Surface Temperature (SST) can increase the Sahelian rainfall during Summer through an increase of moisture transport in the eastern part of the Sahara. He remarked that the rainfall increase is also amplified by a more intense moisture flux from the tropical Atlantic ocean and a more intense local water re-cycling. The Mediterranean Sea may also regulate the northward progress of the African summer monsoon by changing the meridional thermal contrast. The Mediterranean Sea is an important playground for the North Atlantic Oscillation (NAO), a major atmospheric circulation pattern of the Northern Hemisphere. In particular, the Polar–Mediterranean mode, as classified by Kodera and Kuroda (2003), can exert important influences on the Eurasian climate. Yu and Zhou (2004) reported that the cooling trend observed during the recent half century for the subtropical Eurasian continents and for the month of
Regional Atmospheric, Marine Processes and Climate Modelling
377
March is strongly correlated with the DJF (December–January–February) NAO index. They found also that the relation was the strongest with a lag of two months, the time necessary for the cooling signal to propagate from North Africa to Central Asia, in a quasi-barotropic structure for the whole troposphere. The mechanism responsible for this linkage is however still unclear. The existence of large-scale zonally propagated teleconnection structures was already pointed out by Branstator (2002). He demonstrated that the Asian jetstream beginning over North Africa played the role of waveguide by trapping disturbances inside the jet-stream and propagating them from west to east. Watanabe (2004) also found that there is a downstream extension of the NAO during late winter through wavetrain structure. This wavetrain is furthermore interpreted as composed of quasi-stationary Rossby waves trapped on the Asian jet waveguide and excited by the anomalous upper-level convergence over the Mediterranean Sea. He concluded that the Mediterranean convergence associated with the NAO may have some predictability for the medium-range weather forecast in East Asian countries. Li (2005) uses an atmospheric GCM to study the influences of the Mediterranean Sea on the atmosphere. An idealized homogeneous cooling of 2 C for the Mediterranean Sea is imposed as forcing. The model used is the LMDZ, an atmospheric general circulation model with a resolution of 4 in latitude and 5 in longitude. The model was run 9,000 days under perpetual January mode for respectively normal boundary conditions and conditions of an idealized Mediterranean cooling. Figure 125 plots the simulated geopotential height anomalies for 1,000, 850, 500 and 300 hPa respectively. A baroclinic structure is created downstream of the cooling location, across the entire Eurasian continent, roughly following the subtropical jet-stream. There are high (low) pressure anomalies in the lower (upper) atmosphere. Over South Asia, an opposite-sign baroclinic structure is obtained and it is believed to be the consequence of tropical rainfall anomaly. All other remote structures are quasi-barotropic and the most remarkable ones are the deepening of the Aleutian Low in the North Pacific and the weakening of the Icelandic Low in the North Atlantic. In order to study the temporal evolution of the response and the physical mechanisms at different time scales, an ensemble of transient simulations, parallel to the equilibrium runs, are also performed. Each of them lasts 30 days and the ensemble size reaches the huge number of 3,000 to ensure a good statistical significance and an entire coverage of all possible atmospheric states. The approach of the ensemble transient simulations is found very useful in showing the temporal evolution of the response (Li and Conil, 2003). The two teleconnections need several days in the North Pacific and even several tens of days in the North Atlantic to form and to grow. Both of them have a quasi-barotropic
378 Mediterranean Climate Variability
Figure 125: January geopotential height changes (m) at levels of 1,000, 850, 500 and 300-hPa for a homogeneous cooling of 2 C of the Mediterranean sea surface temperature, as simulated in the atmospheric general circulation model LMDZ. vertical structure. It is believed that they are the consequence of complex interactions between the mean flow and the transient eddies in the atmosphere. It is interesting to note that the North Atlantic response is not directly from the Mediterranean Sea, the source of the perturbations, but through a long circle around the world, following roughly the Asian jet-stream and then the North Atlantic sub-polar jet-stream.
Regional Atmospheric, Marine Processes and Climate Modelling
379
7.3. Mediterranean Thermohaline Circulation and its Sensitivity to Atmospheric Forcing The overturning circulation driven by the thermohaline forcing is a particular character of the Mediterranean Sea general circulation. Since the pioneering work of the MEDOC group (MEDOC group, 1970), a large number of studies has been dedicated to water formation in the Mediterranean Sea. An extensive summary can be found in Madec et al. (1991, 1996), Marshall and Schott (1999), Castellari et al. (1998), Lascaratos and Nittis (1998), Lascaratos et al. (1999) and Korres et al. (2000), Beckers et al. (2002) concerning physical mechanisms and numerical modelling. Several factors participate in deep water formation. Firstly, cyclonic structures in the horizontal circulation play an important pre-conditioning role by imposing the dense water in formation to stay at the same place and not to be advected off the formation zone. A second ingredient is the presence of strong atmospheric forcing for both heat flux and wind stress. Convection in the Gulf of Lions and the Adriatic Sea is particularly sensitive to the Mistral and Bora winds which create strong evaporative cooling and wind stress curl when they blow into the sea from the Alps. Intuitively, we can imagine that the performance of the Mediterranean Sea general circulation modelling is quite dependent on the atmospheric forcing and in particular, the intensity of wind stress. This is confirmed by recent experiments performed with the OPA Mediterranean general circulation model at the resolution of 1/8 (MED8, conducted by A. Bozec, unpublished results) and of 1/16 (MED16, conducted by K. Be´ranger, unpublished results). Two datasets of atmospheric forcing are used, one is from ERA40 – the ECMWF re-analysis (T159 model), the other the ECMWF operational analysis (T319). The period used from the re-analysis is from 1990 to 1999 and that from the operational analysis is from August 1998 to August 2002. The ERA run lasts almost 10 years and the ECMWF run is repeated two times to have a simulation of 8 years (in order to have a comparable length with ERA run). It appears that the ERA run is unable to generate convection unless a strong probably unrealistic restoring to winter temperature and salinity is applied, while ECMWF run is able to trigger convection. Figure 126 shows time series of the maximum depth reached by the mixed layer in the Gulf of Lions and in the Levantine basin for the two simulations using MED8 (very similar results are obtained by using MED16). For both Gulf of Lions and Levantine basin, the mixed layer is much deeper in ECMWF run than in ERA run. A stronger interannual variability is also observed in ECMWF run. Although our experimental design does not allow a perfect comparison between the operational analysis and the re-analysis, since they are not over a same time
C
0
0 500
1000
depth (m)
depth (m)
500
1500 2000 JAN
JAN
JAN
1999 2000
2001
JAN
JAN
JAN
2002 1999 2000 time (months)
JAN
2500
JAN
2001 2002
D
0
JAN
JAN
JAN
1999 2000
2001
JAN
JAN
JAN
2002 1999 2000 time (months)
JAN
JAN
2001 2002
0 500
depth (m)
depth (m)
500 1000 1500 2000 2500
1500 2000
2500
B
1000
1000 1500 2000
JAN
JAN
JAN JAN
JAN
JAN JAN
JAN
JAN
1991 1992 1993 1994 1995 1996 1997 1998 1999 time (months)
2500
JAN
JAN
JAN JAN
JAN
JAN JAN
JAN
JAN
1991 1992 1993 1994 1995 1996 1997 1998 1999 time (months)
Figure 126: Time series of the maximum mixed-layer depth (m) in the Gulf of Lion (left panels, A and B) and in the Levantine basin (right panel, C and D) for the model MED8, forced respectively by ECMWF operational analysis (top panels, A and C) and re-analysis ERA40 (bottom panels, B and D).
380 Mediterranean Climate Variability
A
Regional Atmospheric, Marine Processes and Climate Modelling
381
period, we think that the explanation of such large differences is the fact that the ECMWF winds are stronger than those provided by ERA40. A comparison of wind stress for the average of Jan–Feb–Mar between the two datasets is shown in Fig. 127. Although the spatial structure is similar, the intensity in ERA40 seems significantly under-estimated. This is confirmed by S. Marullo (personal communication) who compared both datasets against measurements through wind sensors installed on surface buoys in three locations of the Mediterranean Sea. It is revealed that ECMWF operational analysis winds are quite close to those of the buoy sensors, but the re-analysis winds are under-estimated. The difference in spatial resolution (50 km for ECMWF against 120 km for ERA) is believed to be the main reason to explain the discrepancy of the two datasets. As presented in the following section, the same MED8 model, when forced by Arpege-Climate model stretched to have 50 km for the Mediterranean, did produce marine convection. We may thus generalize the above results and tentatively conclude that the necessary atmospheric resolution is about 50 km in order to simulate the Mediterranean convection and deep water formation.
7.4. Sensitivity of the Mediterranean Thermohaline Circulation to Anthropogenic Global Warming Regional climate changes under global warming context (Jones et al., 1995, 1997; Machenhauer et al., 1998; Frei et al., 2002; Gibelin and De´que´, 2003) are the most important motivations for the Mediterranean regional climate modelling. It is generally agreed that the Mediterranean region is one of the sensitive areas on Earth in the context of global climate change, due to its position at the border of the climatologically determined Hadley cell and the consequent transition character between two very different climate regimes in the North and in the South. In terms of global mean surface air temperature, the Globe has experienced a general warming of 0.6 C over the last century. IPCC (2001) estimated changes of the global temperature to be between 2 to 5 C at the end of the present century. The global mean temperature is only a mean indicator and changes at regional scales can be much larger. Many global and regional models tend to simulate a warming of several degrees (from 3 to 7 C) on the Mediterranean for the end of the twenty-first century and the warming in Summer is larger than the global average. There is also a general trend of a mean precipitation decrease for the region (especially in Summer), due mainly to the northward extension of the descending branch of the subtropical Hadley circulation (IPCC, 2001). Examples and studies concerning the regional projections of global warming are given
45N
A 0.12
4 0.0
Latitude
40N
0.0
4
35N 04
0.
30N 10W
10E 20E Longitude Contour (m2/s): Min= 0.00, Max= 0.17, Int= 0.02 Norme vector (m/s): Min= 0.00, Max= 0.15
30E 0.200000
B 0.04
45N
0E
Latitude
40N 0.04
0.04
35N
0.04
30N 10W
0E
10E 20E Longitude Contour (m2/s): Min= 0.00, Max= 0.09, Int= 0.02 Norme vector (m/s): Min= 0.00, Max= 0.06
30E 0.200000
Figure 127: Wind stress (Pa) for January–February–March, as depicted in ECMWF operational analysis (top) and re-analysis ERA40. The top panel is from 5 years from 1998 to 2002 and the bottom 12 years from 1988 to 1999.
382 Mediterranean Climate Variability
8 0.0
0.0 4
Regional Atmospheric, Marine Processes and Climate Modelling
383
in Chapter 8. Here we only investigate the sensitivity of the Mediterranean thermohaline circulation to global warming. The simultaneous increase of both surface temperature and water deficit (Gibelin and De´que´, 2003; Li, 2003; Giorgi et al., 2004b) could counteract each other in the possible evolution of the Mediterranean Sea thermohaline circulation (MTHC). A weakening or strengthening of the MTHC due to climate change could have an impact on the Mediterranean sea surface temperature and consequently, on the climate of the surrounding areas. With a Mediterranean model at 1/4-degree resolution, Thorpe and Bigg (2000) found that a global warming would lead to a reduced deep water formation in the Mediterranean Sea and to a warmer and saltier outflow. Through the Mediterranean Outflow Waters (MOW), changes of MTHC can furthermore influence the Atlantic Ocean and then the Atlantic thermohaline circulation. The Mediterranean marine ecosystems are also expected to be strongly influenced by the variation of marine circulation. Vichi et al. (2003) investigated the climate change impact on the northern Adriatic Sea and found that an enhanced stratification of the water column, particularly in Summer may reduce the vertical diffusion of oxygen and nutrients. Somot et al. (2005) reported a study employing the Arpege-Climate stretchedgrid model (De´que´ and Piedelievre, 1995; De´que´ et al., 1998; Gibelin and De´que´, 2003) with local spatial resolution around 50 km for the Mediterranean basin. The IPCC-A2 global scenario for the end of the twenty-first century was used. Regional patterns of climate change are similar to those presented in Chapter 8, with a general warming of about 3 C and a decrease of precipitation around the Mediterranean basin. Somot et al. (2005) used furthermore the corresponding changes of atmospheric forcing (wind stress, heat flux, damping SST and water flux) at the sea surface and of river runoff to force a Mediterranean Sea general circulation model at the resolution of 1/8 degree (MED8 model). For the whole Mediterranean Sea and at the end of the twenty-first century, the net heat loss by the surface is lower in the scenario run (1.6 W.m 2) than in the control run (6.1 W.m 2) but the water loss (Evaporation – Precipitation – River runoff ) is higher (0.98 vs. 0.72 m/year). This leads to an increase in temperature and salinity for the Mediterranean Sea (see Table 8) and for each sub-basins. The increase in SST is nearly homogeneous whereas a heterogeneous SSS increase is produced by the model (from þ0.36 psu in the Gulf of Lions to þ0.87 psu in the Aegean Sea). The pattern of SSS anomalies is mainly driven by the river runoff decrease and especially the behaviour of the Po and Black Sea. The competing changes in SST and SSS lead finally to a decrease in surface density and thus a weakening of the MTHC. This weakening is estimated to about 60% for the deep circulation (WMDW: Western Mediterranean Deep Water, EMDW: Eastern Mediterranean Deep Water) and 20% for the
384 Mediterranean Climate Variability
Table 8: Temperature (in C) and salinity (in psu) averaged over different layers of the Mediterranean Sea. ‘‘Control’’ indicates the current climate and ‘‘Scenario’’ at the end of the 21st century.
Control Scenario Diff.
SST
T (0–500m)
T (500-m bottom)
SSS
S (0–500m)
S (500-m bottom)
18.7 21.7 þ3
13.8 15.9 þ2.1
13.0 13.9 þ0.9
38.18 38.61 þ0.43
38.44 38.84 þ0.40
38.66 38.84 þ0.18
intermediate circulation (LIW: Levantine Intermediate Water). The strength of the thermohaline overturning cell can be seen in the Mediterranean zonal overturning stream function (ZOF) following Myers and Haines (2002). The top panel of Fig. 128 plots the ZOF for the control run (30-year average at the end of the control run). The intermediate circulation is seen as a clockwise vertical circulation (positive values) with a maximum value of 1.2 Sv in the Eastern Basin and 1.5 Sv in the Western Basin. This represents mainly the circulations of the Modified Atlantic Water (MAW) and the LIW. The counter-clockwise circulation in the deep part of the Eastern Basin shows the EMDW circulation. A 0.5-Sv circulation is found in the control run. The WMDW path can not be seen by a ZOF. A Western Mediterranean meridional overturning stream function is needed instead. The bottom panel of Fig. 128 plots the ZOF at the end of the scenario simulation (average over the 2070–2099 period). A decrease in the strength and extension of the intermediate thermohaline overturning cell is observed. The deep cell has almost completely vanished. We can thus conclude that the MTHC weakens and becomes shallower during global warming, at least for the IPCC-A2 scenario. Behaviours of the MOW give an integrated mesurement of the Mediterranean Sea evolution. Warmer (þ1.9 C) and saltier (þ0.5 psu) waters are simulated for the end of the twenty-first century. Warm and salty tendencies have also been reported during recent years for the Mediterranean deep waters (Be´thoux et al., 1990; Rohling and Bryden, 1992; Fuda et al., 2002; Rixen et al., 2005) and the MOW (Curry et al., 2003; Potter and Lozier, 2004) from hydrographic data for the last decades. This might be already a manifestation of climate change and global warming for the Mediterranean Sea. The robustness of the results presented by Somot et al. (2005) using only one scenario and one particular model needs, however, to be confirmed by other models and other groups. It will be interesting to explore the validity of the results by incorporating uncertainties in different stages of the investigation
Regional Atmospheric, Marine Processes and Climate Modelling
385
0
-500
-1000
Depth (m)
-1500
-2000
-2500
-3000
-3500
-4000 0
-500
-1000
Depth (m)
-1500
-2000
-2500
-3000
-3500
-4000 -4
0
4
8
12
16
20
24
28
32
36
Longitude
Figure 128: Vertical sections averaged over the 2070–2099 period for Mediterranean Zonal Overturning stream Function for the control run (top) and for the scenario run (bottom). in relation to the emission scenarios, climate projection scenarios, regional downscaling methods and the tunable parameters of the Mediterranean Sea model itself. It will be extremely useful if the multi-model ensemble approach can be performed by different research groups in a coordinated manner. Furthermore, it is worthwhile to note that SST anomalies coming from global coupled model are used for the relaxation of the SST in the regional marine model. This overlooks for example the feedback of the MTHC change to the large-scale SST. As shown by Ra¨isa¨nen et al. (2004) for the Baltic Sea, a more suitable way to overcome this problem is to use an atmosphere–ocean-coupled regional climate model focused on the Mediterranean Basin.
386 Mediterranean Climate Variability
7.5. Current Status of Mediterranean Regional Climate Modelling Global and regional modellings complement each other. While the globalcoupled ocean–atmosphere General Circulation Models (GCM) are the best tools to predict large-scale climate variations at seasonal and interannual scales, and to estimate climate changes at longer time scales, especially those related to the anthropogenic modification of atmospheric composition or surface characteristics, they can not however be directly used in impact-oriented applications because of their relatively coarse spatial scale (typically several hundreds of kilometres). Furthermore, while coarse-resolution coupled GCMs may be capable of capturing the mean climate behaviour, they are usually not successful in reproducing higher order statistics and extreme values. Regional climate modelling has been introduced to fill the gap between the global climate models and the growing demand of climate predictions and scenarios on shorter spatio-temporal scales. Few studies dedicated to the Mediterranean regional climate modelling have been reported so far. Most of the existing research works on climate variability and change over Europe include only partially the Mediterranean basin as the southernmost part of their considered domain. Due to the marginal effects (Giorgi and Francisco, 2000a,b), simulated climates over the Mediterranean basin are often biased by the prescription of the boundary conditions. This decreases the validity of such studies on the Mediterranean climate. One can note however that in Giorgi et al. (2004a,b) and Gibelin and De´que´ (2003), the whole Mediterranean basin is quite in the central part of their regionally oriented studies. The most important regional climate forcing in the Mediterranean region is associated with the complex orography, characterized in many coastal regions by steep mountain slopes, and the large land–sea contrast. These provide a very good testbed but also a big challenge for regional climate modelling. Determination of the Mediterranean regional climate is currently undertaken through several different approaches. The most popular one is the use of (usually atmospheric only) regional climate models (RCM) (Giorgi and Mearns, 1999). The spatial resolution of such models varies from a few kilometres to several tens of kilometres. Models running at resolution less than 10 km are normally based on the full non-hydrostatic equations. Regional climate models (for example, those used in Jones et al., 1995; Christensen et al., 1997; Giorgi and Mearns, 1999; and many others) need to be nested into coarser-resolution global models in order to get the necessary driving information through the lateral boundaries of the domain. This approach allows implementation of highly detailed physical parameterizations in the RCM to ensure a better simulation of local weather and climate events. Another existing approach is based on the use of variable grid
Regional Atmospheric, Marine Processes and Climate Modelling
387
(zoomed) general circulation models (GCM) with higher resolution for the Mediterranean basin (for example, De´que´ and Piedelievre, 1995; Li and Conil, 2003). This ensures a smooth downscaling of information from large scale to regional scale, but the resolution limit is currently thought to be around 50 kilometres, due to limitations in computing capacity and physical parameterizations implemented in such GCMs. A third approach for high-resolution determination of climate parameters over the Mediterranean region is associated with the application of statistical methods for the downscaling of results simulated by large-scale GCMs (Wilby et al., 1998). Several high resolution models of the Mediterranean Sea have been developed during the last decade. These models accurately reproduce the Mediterranean thermohaline circulation and the intermediate and deep water formations which drive it. Among these models, we can mention several 1/8 grid mesh models like OPA (Be´ranger et al., 2004, 2005) and the POM model (Nittis et al., 2003). A 1/16 grid mesh version is also running in the framework of the European Commission-funded programme MFSTEP (Mediterranean Forecasting System: Toward Environmental Predictions) and at IPSL (Be´ranger et al., 2005).
7.6. Atmosphere–Sea Coupled Modelling Obtaining a good representation of the Mediterranean thermohaline circulation is a great challenge for the ocean modelling community because air–sea fluxes need to be simulated with very high accuracy. In the past, many modelling groups were involved in such a challenge, but their oceanic general circulation models were forced by atmospheric fluxes and their sea surface temperatures were relaxed to the observed ones. This relaxation term is a strong constraint for many studies such as the Mediterranean Sea interannual variability and regional climate change projection. Indeed, the impact of the relaxation term on the interannual variability is uncontrolled and often unrealistic. Moreover, in the framework of climate change studies, we do not know how to compute the future SST needed for the surface relaxation. Besides, it is completely impossible, in projecting future scenarios, to take into account the feedback of the evolution of the Mediterranean SST on the local (or global) climate. This justifies the development of an Atmosphere–Ocean Regional Climate Model devoted to Mediterranean studies. The SAMM model (Sea–Atmosphere Mediterranean Model, Sevault et al., 2002) has been developed at CNRM (Centre National de Recherches Me´te´orologiques, Me´te´o-France) coupling the stretched version of ArpegeClimate (De´que´ and Piedelievre, 1995) and MED8 as used in the previous
388 Mediterranean Climate Variability
sections. The atmospheric component has a horizontal resolution of 50 km over the Mediterranean Basin and the Mediterranean Sea component has a resolution of about 10 km. Each day, the two components exchange SST as well as momentum, water and heat fluxes. At the surface, the interaction between the Mediterranean Sea and the atmosphere is completely free because the simulation has been run without relaxation or flux correction. For the river runoff fluxes, a monthly climatology is computed from the RivDis database (Vo¨ro¨smarty et al., 1996). Specific parameterizations are used for the Black Sea (based on salt conservation) and for the Nile (in order to obtain realistic runoff for the period after the building of the Aswan dam). Outside the Mediterranean Sea, the SST used in the atmospheric model is prescribed from interannual monthly mean observed data, reconstructed with in situ and satellite data (Smith et al., 1996). A 38-year simulation has been performed with SAMM following a 20-year spin-up. The area and the coast line of the model are presented in Fig. 129 as well as the winter averaged 34 m-depth temperatures and horizontal currents. For comparison purposes, a parallel experiment has been carried out with MED8 forced by air–sea fluxes coming from a previously run using only the atmospheric Arpege-Climate model. The only difference between the simulations is thus the way of taking into account the air–sea fluxes, which permits one to quantify the differences between fully coupled and uncoupled models. The surface water flux (Evaporation–Precipitation) for the SAMM simulation and over the whole basin is equal to 0.77 m/year with a weak standard deviation in agreement with observed evidence. Note that the river runoff flux is prescribed according to its seasonally-varied obervation-based estimation with an annual average of 0.18 m/year. The same computation for the surface net heat flux gives a value of 7.1 W/m2 (heat loss for the Mediterranean Sea) with a standard deviation of 5.0 W/m2. These values are in agreement with observed data and other modelling studies. In SAMM as in the real world, the surface heat loss is compensated by a positive heat transport across the Gibraltar Strait (þ5.5 W/m2, with a weak standard deviation of 0.3 W/m2). Note that the values are normalized by the surface of the Mediterranean area to be consistent with the surface flux. The small negative total heat budget, 1.6 W/m2 ( 7.1 þ 5.5 W/m2) implies a weak cooling drift occurring along the simulation. But it is not statistically different from zero due to the large interannual variability of the heat content change (standard deviation of 5.1 W/m2). The time series of the net surface heat flux, the Gibraltar heat transport and the heat content change are plotted in Fig. 130 for the coupled simulation (left panel) and for the forced simulation (right panel). The time correlation between the surface flux and the heat content is equal to 0.98 for both simulations. The comparison of the interannual variability of these 3 terms implies that all the surface flux variability is damped by the heat
45N
Latitude 35N
30N 10W
10.00
0E
11.00
20E
10E Longitude
12.00
30E 0.4 m/s
13.00
14.00
15.00
Figure 129: Winter (January, February, March) 34-m depth horizontal currents (arrow, in m/s) and potential temperature (colour, in C), averaged over the whole coupled simulation (38 years).
Regional Atmospheric, Marine Processes and Climate Modelling
40N
389
390 Mediterranean Climate Variability
20
15
10
5
0
−5
−10
−15
−20 1960
1965
1970
1975
1980
1985
1990
1995
1960
1965
1970
1975
1980
1985
1990
1995
Figure 130: Time evolution of the Mediterranean Sea heat budget components (in W/m2) computed from the coupled simulation (left) and from the forced simulation (right). The surface heat flux is plotted in solid line, the Gibraltar heat transport in dashed line and the heat content change term in dotted line.
content of the Mediterranean Sea and not exported across the Gibraltar Strait. Indeed, the physical constraints due to the shape of the strait lead to a filtering of the interannual variability. Another interesting feature is that the interannual variability (standard deviation) simulated in the coupled model is always lower than in the forced model. Even if the simulations are not long enough to obtain statistically significant results, this variability difference is obtained for many variables, both globally and locally. For example, this is true for the surface heat flux and the heat content averaged over the Mediterranean Sea but also in the Gulf of Lions area and in the Adriatic Sea. For these two sub-basins of deep water formation, a lower interannual variability in the coupled model is also observed for the water mass formation rate and the deep water volume transport. Further work is needed to better understand this behaviour but the coupled model seems to simulate an additional air–sea feedback which is not represented in the forced ocean model. We are thus convinced that regional coupled models are much more suitable to study physical mechanisms and climate interannual variability, and to make future projections of regional climate change (see also Section 7.7.2).
7.7. Perspectives and Outlooks In this chapter, we have shown several examples where numerical modelling was used to investigate physical mechanisms controlling the Mediterranean
Regional Atmospheric, Marine Processes and Climate Modelling
391
climate variation and change. Due to its particular geographical position, the Mediterranean region is quite strongly related to other major climate phenomena of the globe, such as tropical monsoons and the North Atlantic Oscillation. The Mediterranean Sea can also exert its climatic influence on the nearby and remote regions in Africa, Europe and Asia through complex atmospheric processes. It is also believed that the Mediterranean outflow water plays an important role for the Atlantic overturning circulation and ultimately the global climate. Under the global warming context, current atmospheric models seem to converge on the conclusion that both water stress on the nearby lands and water deficit of the Mediterranean Sea itself increase. This may further impact the marine overturning circulation and the marine ecosystem. Considering the results reviewed in this chapter, two important issues can be foreseen for the Mediterranean regional climate modelling in the next few years.
7.7.1. High-Resolution Mediterranean Climate Modelling Systems The spatial resolution of future modelling systems will be further increased. It is expected to have regional atmospheric models with resolution around 10 to 20 kilometres in the next few years. Experience with numerical weather forecasting shows that higher spatial resolution usually leads to better prediction, mainly due to improvements in the representation of atmospheric instability which is crucially dependent on the model’s spatial resolution. In climate modelling, higher spatial resolution may lead to improvements in some aspects and degradation in others (May and Roeckner, 2001; Leung et al., 2003). Climate is in fact more related to the sources and sinks of energy, moisture and momentum. Mechanisms controlling their budgets and evolution at different spatio-temporal scales are thus crucial for climate. In general higher spatial resolution models can provide a more comfortable background to incorporate sophisticated physics and the latter will improve the performance of regional climate models. For the Mediterranean region, high resolution is particularly important, as shown in Section 7.4, since there is a very complex terrain surrounding the Mediterranean Sea, responsible for intense wind events, such as Mistral and Bora which contribute largely to oceanic convection in the Mediterranean (Gulf of Lions, Adriatic Sea and Aegean Sea). The overall studies reported in the current scientific literature seem to show improved model performance with higher spatial resolution, especially in reproducing extreme events, such as strong precipitation episodes and cyclogenesis often related to the specific surface orography. But there is indeed a need to further evaluate and quantify the impacts of spatial resolution on regional climate simulation. Even in the most advanced high-resolution regional climate models,
392 Mediterranean Climate Variability
it will be difficult, in some cases, to determine dynamically the hydrological variables, such as run-off. Application of statistical methods will always be necessary to provide appropriate solutions for climate change impact studies. In the next few years, high-resolution Mediterranean climate modelling systems are expected to be used to produce consistent data for the Mediterranean basin during the last 40 years, which can not be achieved by global re-analysis performed at weather prediction centres (such as NCEP and ERA40) due to the too coarse spatial resolution and the deficiency in the hydrological cycle. By performing a special calibration through the regional atmospheric/land-surface climate models covering a quite large domain around the Mediterranean, it is in principle possible to reduce the hydrological bias of the re-analysis products. Such simulations of the Mediterranean climate over the last 40 years will be very useful to study the dynamical and physical processes controlling the climate in the Mediterranean region. They are also useful for climate trend detection for the last 40 years.
7.7.2. Development and Validation of Integrated Regional Modelling Systems Other components controlling the regional climate will enter interactively into the regional modelling system. They include, through the most important topics, the Mediterranean Sea general circulation, basin-scale hydrology, dynamic surface vegetation, land use, atmospheric chemistry, air pollution and manmade or desert-originated aerosols, marine and land-surface ecosystems. It is expected that new climate feedbacks and modes derived from the complex interaction among different components of the Mediterranean climate system might be discovered and quantified. Especially the regional atmosphere and Mediterranean Sea coupled models should receive high priority for their development and utilisation in the Mediterranean climate studies. With increasing complexity of numerical modelling systems, validation against appropriate observational data is becoming an important issue. This will require however a significant improvement of the currently existing data bases for the region and an increasing capacity to obtain and analyse new measurements with different geophysical characteristics of the region like soil moisture, soil types, vegetation coverage, dust sources and transport, etc. The current observational network around the Mediterranean basin is still scarce and accuracy of measured geophysical parameters in this region is also significantly lower than that over more developed areas like Europe. Special emphasis will be made on the processing of satellite data dedicated to measure surface processes such as sea surface temperature and height, and vegetation. Initiatives as those managed by CIESM to monitor deep sea hydrology will be encouraged as they provide mandatory controls for the climate models.
Regional Atmospheric, Marine Processes and Climate Modelling
393
Putting the numerical systems in the configuration of paleoclimate will be an interesting exercise to test the robustness of the numerical models because it is the only way to test the sensitivity of our complex models to documented climate changes. Paleoclimate simulations will allow to test not only the ability of models to simulate the correct amplitude but also the geographical pattern of climate changes thanks to a large number of dated samples all around the Mediterranean basin. It should be noted that climate studies on these timescales require also outputs from global general circulation models.
Acknowledgements This work is supported by the French national programme GICC (Gestion et Impact du Changement Climatique). Many people have contributed to the present paper or its earlier versions: S. Hagemann, D. Jacob, R. Jones, E. Kaas, S. Krichak, P. Lionello, A. Mariotti, B. Weare, among others.
References Alhammoud, B., Be´ranger, K., Mortier, L., & Cre´pon, M. (2005). Surface circulation of the Levantine Basin: comparison of model results with observations. Progress in Oceanography, 66, 299–320. Beckers, J. M., & 23 co-authors (2002). Model intercomparison in the Mediterranean: MEDMEX simulation of the seasonal cycle. J. of Marine Systems, 33–34, 215–251. Be´ranger, K., Mortier, L., Gasparini, G. P., Gervasio, L., Astraldi, M., & Cre´pon, M. (2004). The dynamics of the Sicily Strait: a comprehensive study from observations and models. Deep-sea Reasearch II, 51, 411–440. Be´ranger, K., Mortier, L., & Cre´pon, M. (2005). Seasonal variability of water transport through the Straits of Gibraltar, Sicily and Corsica, derived from a high-resolution model of the Mediterranean circulation. Progress in Oceanography, 66, 341–364. Bethoux, J. P., Gentili, B., Raunet, J., & Tailliez, D. (1990). Warming trend in the Western Mediterranean deep water. Nature, 347, 660–662. Branstator, G. (2002). Circumglobal teleconnections, the jet stream waveguide, and the north atlantic oscillation. J. of Climate, 15, 1893–1910. Castellari, S. N., Pinardi, & Leaman, K. D. (1998). A model study of air-sea interactions in the Mediterranean Sea. J. of Marine Systems, 18, 89–114. Christensen, J. H., Machenhauer, B., Jones, R. G., Scha¨r, C., Ruti, P. M., Castro, M., & Visconti, G. (1997). Validation of present-day climate simulations over Europe: LAM simulations with observed boundary conditions. Climate Dynamics, 13, 489–506. Curry, R., Dickson, B., & Yashayaev, I. (2003). A change in the freshwater balance of the Atlantic Ocean over the past four decades. Nature, 426(6968), 826–828. De´que´, M., & Piedelievre, J. P. (1995). High-resolution climate simulation over Europe. Climate Dynamics, 11, 321–339.
394 Mediterranean Climate Variability
De´que´, M., Marquet, P., & Jones, R. G. (1998). Simulation of climate change over Europe using a global variable resolution general circulation model. Climate Dynamics, 14, 173–189. Frei, C., Christensen, J. H., De´que´, M., Jacob, D., Jones, R. G., & Vidale, P. L. (2002). Daily precipitation statistics in regional climate models: evaluation and intercomparison for the European Alps. J. of Geophysical Research, 108(D3), 4124–4142. Fuda, J. L., Etiope, G., Millot, C., Favali, P., Calcara, M., Smirglio, G., & Boschi, E. (2002). Warming, salting and origin of the Tyrrhenian Deep Water. Geophysical Research Letters, 29(19), 1898, doi:10.1029/2001GL014072. Gascard, J. C., & Richez, C. (1985). Water masses and circulation in the western Alboran Sea and in the Strait of Gibraltar. Progress in Oceanography, 15(3), 157–216. Gibelin, A. L., & De´que´, M. (2003). Anthropogenic climate change over the Mediterranean region simulated by a global variable resolution model. Climate Dynamics, 20, 327–339. Giorgi, F., & Francisco, R. (2000a). Evaluating uncertainties in the predictioon of regional climate change. Geophysical Research Letters, 27, 1295–1298. Giorgi, F., & Francisco, R. (2000b). Uncertainties in regional climate prediction: a regional analysis of ensemble simulations with the HADCM2 coupled AOGCM. Climate Dynamics, 16, 169–182. Giorgi, F., & Mearns, L. O. (1999). Introduction to special section: regional climate modelling revisited. J. of Geophysical Research, 104, 6335–6352. Giorgi, F., Bi, X., & Pal, J. S. (2004a). Mean, interannual variability and trends in a regional climate change experiment over Europe. I. present-day climate (1961–1990). Climate Dynamics, 22, 733–756. Giorgi, F., Bi, X., & Pal, J. (2004b). Mean, interannual variability and trends in a regional climate change experiment over Europe. II. climate change scenarios (2071–2100). Climate Dynamics, 23, 839–858. Hamad, N., Millot, C., & Taupier-Letage, I. (2002). The surface circulation in the eastern basin of the Mediterranean Sea: new elements, in: Proceedings of the 2nd International Conference on Oceanography of the eastern Mediterranean and Black Sea: similarities and difference of two interconnected basins, Ankara, Turkey 14–18 October, 2002. Heburn, G. W., & La Violette, P. E. (1990). Variations in the structure of the anticyclonic gyres found in the Alboran Sea. J. of Geophysical Research, 95, 1599–1613. Herbaut, C., Codron, F., & Cre´pon, M. (1998). Separation of a costal current at a strait level: case of the Strait of Sicily. J. of Physical Oceanography, 28, 1346–1362. Jones, R. G., Murphy, J. M., & Noguer, M. (1995). Simulation of climate change over Europe using a nested regional-climate model: I: assessment of control climate, including sensitivity to location of lateral boundaries. Quarterly Journal of the Royal Meteorological Society, 121, 1413–1449. Jones, R. G., Murphy, J. M., Noguer, M., & Keen, A. B. (1997). Simulation of climate change over Europe using a nested regional-climate model. II: comparison of driving and regional model responses to a doubling of carbon dioxide. Quarterly Journal of the Royal Meteorological Society, 123, 265–292. Kodera, K., & Kuroda, Y. (2003). Regional and hemispheric circulation patterns in the Northern Hemisphere winter, or the NAO and the AO. Geophysical Research Letters, 30, 1934, doi:10.1029/2003GL017290.
Regional Atmospheric, Marine Processes and Climate Modelling
395
Korres, G., Pinardi, N., & Lascaratos, A. (2000). The ocean response to low-frequency interannual atmospheric variability in the mediterranean Sea. Part I: sensitivity experiments and energy analysis. J. of Climate, 705–731. Lascaratos, A., & Nittis, K. (1998). A high-resolution three-dimensional numerical study of Intermediate Water formation in the Levantine Sea. J. of Geophysical Research, 103(C9), 18497–18511. Lascaratos, A., Roether, W., Nittis, K., & Klein, B. (1999). Recent changes in deep water formation and spreading in the Eastern Mediterranean Sea: a review. Progress in Oceanography, 44, 5–36. Leung, L. R., Means, L. O., Giogi, F., & Wilby, R. L. (2003). Regional climate research, needs and opportunities. Bulletin of the American Meteorological Society, 84, 89–95. Lermusiaux, P. F. J., & Robinson, A. R. (2001). Features of dominant mesoscale variability, circulation patterns and dynamics in the Strait of Sicily. Deep-sea reasearch I, 48, 1953–1997. Li, L. (2003). Evolution future du climat en Me´diterrane´e: vers un e´tat de se´cheresse accru? Rapport CNFGG 2003, 220–223, Paris, France. Also available at http:// www.omp.obs-mip.fr/cnfgg/UGGI2003.pdf. Li, Z. X. (2005). Atmospheric GCM response to an idealized anomaly of the Mediterranean Sea surface temperature. Clim. Dyn., submitted. Li, Z. X., & Conil, S. (2003). Transient response of an atmospheric GCM to North Atlantic SST anomalies. J. of Climate, 16, 3993–3998. Machenhauer, B., Windelband, M., Botzet, M., Christensen, J. H., De´que´ M., Jones, R. G., Ruti, P. M., & Visconti, G. (1998). Validation and analysis of regional presentday climate and climate change simulations over Europe. MPI Report No. 275, MPI, Hamburg, Germany. Madec, G., Chartier, M., Delecluse, P., & Cre´pon, M. (1991). A three-dimensional numerical study of deep-water formation in the northwestern Mediterranean Sea. J. of Physical Oceanography, 21, 1349–1371. Madec, G., Lott, F., Delecluse, P., & Cre´pon, M. (1996). Large-scale preconditioning of deep-water formation in the Western Mediterranean Sea. J. of Physical Oceanography, 26, 1393–1408. Mariotti, A., Struglia, M. V., Zeng, N., & Lau, K. M. (2002). The hydrological cycle in the Mediterranean region and implications for the water budget of the Mediterranean Sea. J. of Climate, 15, 1674–1690. Marshall, J., & Schott, F. (1999). Open-ocean convection: observations, theory and models. Reviews of Geophysics, 37(1), 1–64. Marullo, S., Santoleri, R., Malanotte-Rizzoli, P., & Bergamasco, A. (1999). Atlantic water in the Strait of Sicily. J. of Geophysical Research, 95, 1569–1575. May, W., & Roeckner, E. (2001). A time-slice experiment with the ECHAM4 AGCM at high horizontal resolution: the impact of the horizontal resolution on annual mean climatic change. Climate Dynamics, 17, 407–420. Mearns, L. O., Bogardi, I., Giorgi, F., Matyasovszky, I., & Palecki, M. (1999). Comparison of climate change scenarios generated from regional climate model experiments and statistical downscaling. J. of Geophysical Research, 104, 6603–6621. MEDOC Group (1970). Observation of formation of deep water in the Mediterranean in 1969. Nature, 227, 1037–1040.
396 Mediterranean Climate Variability
Millan, M. M., & 21 co-authors (2005a). Climatic feedbacks and desertification: the mediterranean model. J. of Climate, 18, 684–701. Millan, M. M., Estrela, M. J., & Miro, J. J. (2005b). Rainfall components: variability and spatial distribution in a mediterranean area (Valencia region). J. of Climate, 18, 2682–2705. Millot, C. (1987). Circulation in the Western Mediterranean Sea. Oceanologica Acta, 10, 143–149. Myers, P., & Haines, K. (2002). Stability of the Mediterranean’s thermohaline circulation under modified surface evaporative fluxes. J. of Geophysical Research, 107(C3), 3021, doi:10.1029/2000JC000550. Nittis, K., Lascaratos, A., & Theocharis, A. (2003). Dense water formation in the Aegean Sea: numerical simulations during the Eastern Mediterranean Transient. Journal of Geophysical Research, 108(C9), 8120, doi:10.1029/2002JC001352. Potter, R., & Lozier, S. (2004). On the warming and salinification of the Mediterranean outflow waters in the North Atlantic. Geophysical Research Letters, 31, 01202, doi:10.1029/2003GL018161. Ra¨isa¨nen, J., Hansson, U., Ullerstig, A., Do¨scher, R., Graham, L. P., Jones, C., Meier, H. E. M., Samuelsson, P., & Wille´n, U. (2004). European climate in the late twenty-first century: regional simulations with two driving global models and two forcing scenarios. Climate Dynamics, 22, 13–31. Rixen, M., & 16 co-authors (2005). The Western Mediterranean Deep water: a proxy for climate change. Geophysical Research Letters, 32, L12608, doi: 10.1029/2005GL022702. Robinson, A. R., Sellschopp, J., Warn-Varnas, A., Leslie, W. G., Lozano, C. J. Jr., Haley, P. J., Anderson, L. A., & Lermusiaux, P. F. J. (1999). The Atlantic Ionian stream. J. of Marine Systems, 20, 129–156. Rohling, E., & Bryden, H. (1992). Man-induced salinity and temperature increases in the Western Mediterranean deep water. J. of Geophysical Research, 97, 11191–11198. Rowell, D. P. (2003). The impact of Mediterranean SSTs on the Sahelian rainfall season. J. of Climate, 16, 849–862. Schar, C., Vidale, P. L., Luthi, D., Frei, C., Haberli, C., Liniger, M. A., & Appenzeller, C. (2004). The role of increasing temperature variability in European summer heatwaves. Nature, 427, 332–336. Sevault, F., Somot, S., & De´que´, M. (2002). Couplage ARPEGE-MEDIAS -OPAMEDITERRANEE. Les e´tapes, Note de Centre, CNRM, GMGEC, n. 84, MeteoFrance, Toulouse, France (in French, available at
[email protected]). Smith, T., Reynolds, R., Livezey, R., & Stokes, D. (1996). Reconstruction of historical sea surface temperatures using empirical orthogonal functions. J. of Climate, 9, 1403–1420. Somot, S., Sevault, F., & De´que´, M. (2005). Is the Mediterranean Sea thermohaline circulation stable in a climate change scenario? Climate Dynamics, in revision. Struglia, M. V., Mariotti, A., & Filograsso, A. (2004). River discharge in the Mediterranean Sea: climatology and aspects of the observed variability. J. of Climate, 17, 4740–4750. Thorpe, R. B., & Bigg, G. R. (2000). Modelling the sensitivity of Mediterranean outflow to anthropogenically forced climate change. Climate Dynamics, 16, 355–368. Vichi, M., May, W., & Navarra, A. (2003). Response of a complex ecosystem model of the northern Adriatic Sea to a regional climate change scenario. Climate Research, 24, 141–159.
Regional Atmospheric, Marine Processes and Climate Modelling
397
Vo¨ro¨smarty, C., Fekete, B., & Tucker, B. (1996). Global river discharge database, RivDis, vol. 0 to 7. International Hydrological Program, Global Hydrological Archive and Analysis Systems, UNESCO, Paris, France. Watanabe, M. (2004). Asian jet waveguide and a downstream extension of the North Atlantic Oscillation. J. of Climate, 17, 4674–4691. Wilby, R. L., Wigley, T. M. L., Conway, D., Jones, P. D., Hewitson, B. C., Main, J., & Wilks, D. S. (1998). Statistical downscaling of general circulation model output: A comparison of methods. Water Resources Research, 34, 2995–3008. Yu, R., & Zhou, T. (2004). Impacts of winter-NAO on March cooling trends over subtropical Eurasia continent in the recent half century. Geophysical Research Letters, 31, 12204, doi: 10.1029/2004GL019814.
This page intentionally left blank
Chapter 8
The Mediterranean Climate Change under Global Warming U. Ulbrich,1 W. May,2 L. Li,3 P. Lionello,4 J. G. Pinto5 and S. Somot6 1
Freie Universita¨t Berlin, Germany (
[email protected]) Danish Meteorological Institute, Denmark (
[email protected]) 3 LMD/IPSL/CNRS, Universite´ P. et M. Curie, Paris, France (
[email protected]) 4 University of Lecce, Italy (
[email protected]) 5 Universita¨t zu Ko¨ln, Germany (
[email protected]) 6 Me´te´o-France, CNRM/GMGEC/EAC Toulouse, France (
[email protected]) 2
8.1. Introduction Surprisingly, the issue of climatic change in the Mediterranean region has rather sparsely been addressed in studies performed more than 5–10 years ago. The region has recently received increasing scientific interest. In fact, the Mediterranean has some particular characteristics that demand to put it high on the research agenda. Not only is the Mediterranean connected with many other parts of the globe by long-range atmospheric teleconnections, but also changes of the inflow from the Mediterranean Sea into the North Atlantic affect the thermohaline circulation in the Atlantic (see Chapter 7). Further, the Mediterranean region has a very specific climate with very dry summers and wet winters, which may be very sensitive to potential climatic changes. In the 1990s, several studies have used simulations from global-coupled climate models in order to assess the potential climatic change in Europe. Only a few of these studies considered the climatic change in the Mediterranean region in particular. Some of them applied statistical downscaling techniques, others analysed the output from global-coupled climate models directly. Recently, multi-model GCM analyses have looked at the consistency of climatic change signals across simulations and scenarios, and one of the regions considered in an integral way is the Mediterranean.
400 Mediterranean Climate Variability
The coarse resolution of the global-coupled climate models implies, however, certain limitations. Regional details of the climate in the Mediterranean region cannot be simulated realistically. A prime reason is the missing of important regional characteristics such as the complex mountain ranges and the distribution of land. Further, these models have shortcomings in simulating Mediterranean cyclones and regional phenomena such as heavy rainfall events realistically. One way to overcome these problems is by using Regional Climate Models (RCMs), driven by information provided by a global-coupled model at its lateral and lower boundaries. There was however, only one study which applied a RCM for the entire Mediterranean region, while many other simulations only included the northern and western parts of the Mediterranean. Other options are global Atmospheric General Circulation models (AGCMs) with high horizontal resolution, either with the same resolution globally or with high horizontal resolution in the Mediterranean region and lower resolution elsewhere. In this chapter, an overview of the scientific literature on climatic change in the Mediterranean region due to the anticipated global warming is given. We distinguish between the results obtained from global-coupled climate models, results obtained via regionalization techniques (i.e. statistical and dynamical downscaling) and results obtained from global high-resolution AGCMs. We also comment on the impacts of climate change on the land surface and discuss the needs of future research on climate change in the Mediterranean region.
8.2. Global-Coupled Climate Models Different model simulations of anthropogenic climate change can lead to different estimates of climatic changes for a number of reasons. GCMs have, for instance, different parameterizations of subscale processes. This may lead to differences in the simulated climatic feedback mechanisms, eventually affecting the climatic change signals produced. Another aspect is the existence of decadal climate variations that one may imagine as being superposed to a greenhouse gas signal. They may play an important role when rather short simulation episodes are considered. Further, different scenarios for various greenhouse gases and aerosols can be used. These effects are very complex and may lead to some of the differences between the studies referred to in the following. As for the mean temperatures at 2-m height, the common future change is an increase both in winter and summer. Typical ranges are 2–4 K with CO2 doubling for both summer and winter (De´que´ et al., 1998). This trend was confirmed by the ACACIA experiments constructed from 5 different global-coupled models
The Mediterranean Climate Change under Global Warming
401
(Parry, 2000). They found an increase of the winter temperatures from the 1961– 1990 mean to the 2080s which is in the range of 4–5 K (6–7 K) in winter (summer) using the A2 marker scenario (Houghton et al., 2001). Some simulations of the greenhouse gas-induced climate change show a northward shift of the North Atlantic winter storm track going along with a shift and intensification of the North Atlantic Oscillation (Ulbrich and Christoph, 1999). This leads to increasing precipitation in northern Europe but reductions over many parts of the Mediterranean. A close link between winter rainfall and baroclinic activity over the North Atlantic has been demonstrated by Ulbrich et al. (1999) for Portugal and by Knippertz et al. (2003) for north western Morocco. Thus, both the northward shift of baroclinic waves and reduced moisture transports from the Atlantic are consistent with the reduced precipitation in the scenario simulations. The ECHAM4 model simulation following the IS92a scenario, for instance, shows a strong decrease of precipitation over the whole Mediterranean basin, in particular over the western Iberian peninsula, southern Turkey, the Near East, and Egypt, with the changes exceeding 30% (Fig. 131). These results are consistent with Sanchez et al. (2004), considering a simulation with a different climate model. In contrast, De´que´ et al. (1998) found 30% increases in winter precipitation over the Mediterranean with a doubling of CO2 concentrations using the ARPEGE model. Also the ACACIA scenarios (Parry, 2000) are more in line with the latter study, suggesting an increase (decrease) of winter rainfall in the northwestern (southeastern) Mediterranean basin towards the 2080s. For summer, all models produce a reduction of rainfall, typically on the order of 10–30%. The ACACIA simulations, on the other hand, give a reduction of about 50% in summer rainfall (Parry, 2000). Multi-model GCM analyses have looked at signals in different regions of the world, including the Mediterranean (Kittel et al., 1998; Giorgi and Francisco, 2000; Giorgi et al., 2001a,b). These studies suggest that summer drying over the Mediterranean is a consistent signal across different GCMs and for different scenarios. Concerning the winter results, it is elucidating to consider the origin of the different signals, for example looking into the statistics of the simulated cyclones. Due to their small scale, the analysis of Mediterranean cyclones in scenario simulations is a difficult task. Results based on the ECHAM4 simulation (Pinto et al., 2005a,b) show a rather realistic representation of cyclonic activity over the basin (Fig. 132A), although the details of the very small-scale cyclones could not be resolved due to the coarse model resolution (T42). The changes in cyclonic activity in a 2 CO2 situation are characterized by a strong reduction of the number of cyclones over the basin (Fig. 132B) and by a general northward shift of the cyclone tracks. These are consistent with the changes in precipitation mentioned in the preceding paragraph and can largely be attributed to alterations
402 Mediterranean Climate Variability
Figure 131: Difference of annual precipitation related to greenhouse gas forcing (scenario IS92a, difference of simulated rainfall between 50 year averages for years 2039–2089 and 1880–1930) as produced by the ECHAM4/OPYC3 coupled atmosphere–ocean GCM. Units (colour bar) are mm. The relative differences (in %) are given by hatching.
in the sea-level pressure and the upper tropospheric baroclinicity (Figs. 132C, D), showing less favourable conditions for the development of cyclones at lower latitudes, including the Mediterranean region. In another study, the differences in cyclonic activity have been estimated for two 30-year long-time slice experiments, carried out with the ECHAM4 model at a high resolution of T106 (Lionello et al., 2002). For the present day climate, this model version shows a slightly higher number of cyclones as compared to the low-resolution model (Fig. 133, left panel). The climate change signal is a decrease in cyclone number. The doubled CO2 simulation is
B
C
D
403
Figure 132: (A) Winter (Oct–Mar) cyclone track density (cyclone days/winter) for the present day climate in the ECHAM4/OPYC3 atmosphere–ocean GCM; (B) As (A) but differences between a climate with increased greenhouse gas forcing (mid-21st century, forcing according to the IS92a scenario) and present climate; (C) as Fig. (B) but for mean SLP (hPa) and (D) as Fig. (B) but for upper tropospheric baroclinicity (day 1). For all panels, area with statistically significant differences (99% level according to a t test) are shaded.
The Mediterranean Climate Change under Global Warming
A
N/Year
404 Mediterranean Climate Variability
40
2.0
30
1.5
20
1.0
10
.5
0 15
20
25
30
35
.0 35
40
45
50
55
Depth (hpa)
Figure 133: Left panel: Cumulated distribution, that is the average number of cyclones per year (y axis) whose depth exceeds a given threshold (x axis, in hPa) for the present scenario (dotted line), doubles CO2 (dashed line) and ERA-15 (solid line) scenarios; Right panel: The tail of the cumulated distributions, shown in the left panel, that is cyclones with depth exceeding 35 hPa (from Lionello et al., 2002).
characterized by more deep lows, but the difference between the two time-slices is hardly significant (Fig. 133, right panel). The changes are not associated with a variation of the regions of formation of the cyclones. A recent study on the climatic change over Europe based on several coupled climate models has shown a consistent reduction of cyclonic activity for the Mediterranean region under future climate conditions (Leckebusch et al., 2005). The same tendency is further confirmed by Somot (2005) considering a GCM with variable resolution and employing a different identification and tracking scheme for the cyclones (see Section 8.3.3). One of the effects imposed by cyclones is the occurrence of strong winds and their effects. Using the 30-year time-slice experiments mentioned above, Lionello et al. (2003) investigated future changes in storm surges in the Adriatic Sea, using a dynamical downscaling approach, but did not find any statistically significant change in the extreme surge level. On the other hand, a reduction of the extreme wave height in a doubled CO2 scenario was found in the southern Adriatic Sea. In a recent extension of this work, Lionello et al. (2006) used wind fields computed by regional climate simulations to compute the wave field. Figure 134 shows the mean annual cycle for the present climate (CTR) and those corresponding to the A2 and B2 emission scenarios, respectively. The most pronounced signal is the reduction of wave height in November, December and May. In this section, we have restricted ourselves to a discussion on some simulated effects at the ocean surface. Work on the sensitivity of the Mediterranean thermohaline circulation to anthropogenic global warming is described in Section 7.5 of this book.
The Mediterranean Climate Change under Global Warming
405
Figure 134: Mean annual cycle of monthly surge wave height (values in m, y axis) in the A2 (grey curve), B2 (dashed curve) and present climate (CTR, black curve) simulations. Calendar month on the x axis. The vertical bars show the standard deviation of the mean monthly values (from Lionello et al., 2006).
8.3. Regional Climate Scenarios 8.3.1. Statistical Downscaling The coupled climate models currently used for simulating climate change have a coarse spatial resolution of several 100 km. They are generally able to reproduce large-scale circulation characteristics, but not the regional climate parameters. A common approach used for obtaining local values is statistical downscaling. While this approach has frequently been applied in studies dealing with present day climate, its use for estimations of future climate in the Mediterranean region has been rather limited. Trigo and Palutikof (2001) and Gonza´lez-Rouco et al. (2000) have applied several statistical downscaling methods in order to obtain precipitation scenarios over Iberian peninsula. They used the HadCM2 greenhouse gas plus sulphate scenario simulations as a basis. As for precipitation over the Iberian peninsula, they found an increase of precipitation in winter and small decreases for spring and autumn.
406 Mediterranean Climate Variability
Sumner et al. (2003) also considered precipitation in Spain, but based their study on the ECHAM4/OPYC3 output. Comparing the time periods 1971–90 and 2080–99, they used the closest analogue amongst 19 identified atmospheric circulation patterns under present-day conditions for downscaling. This approach, however, assumes that the relationships between circulation type and daily precipitation distribution remain the same also for the future climate. The statistical model produced decreases in the frequency of circulations with a westerly or northerly component. Their conclusion is that the changes in circulation lead to opposing signals in different regions. They inferred a reduction in annual precipitation for Andalusia and the upland parts of Catalonia (6–14%), while an increase of up to 14% was found along the northern part of the Spanish Mediterranean coast.
8.3.2. RCM Simulations In the PRUDENCE project, several RCMs have been run with lateral boundary conditions from different global General Circulation Model. Comparing two 30-year time-slices (1961–1990 for present day, 2071–2100 from the A2 scenario), Ra¨isa¨nen et al. (2004) found widespread increases of the mean temperatures in the Mediterranean region, amounting 6 K and more in summer over the land areas and less pronounced increases in winter (3–4 K). Differences between the results from two different driving models (ECHAM4 and HadAM3H) investigated by Ra¨isa¨nen et al. (2004) are much smaller than the mean signal. Sanchez et al. (2004) considered maximum daily temperatures in a regional model driven by HadAM3H, finding changes very similar to those obtained by Ra¨isa¨nen et al. (2004) for mean temperatures. It is interesting to note that the number of heat waves (temperature anomalies exceeding 5 K occurring over more than 6 consecutive days) are not increasing in all regions. Sanchez et al. (2004) interpreted this result in terms of a reduced duration of heat waves at some locations, while the number of extremely hot days increases. Further, it was noted that minimum temperatures also increase, with stronger changes in summer (about 5 K) than in winter (3 K). Sanchez et al. (2004) and Ra¨isa¨nen et al. (2004) also computed average precipitation changes, finding different signs of the signal (depending on the season and the particular part of the Mediterranean region). The simulations produce rainfall increases of up to about 20% in the western and northern parts of the Mediterranean in winter when driven by HadAM3H but reductions in the southern parts (about 10%). Using ECHAM4, the region with increased precipitation is shifted northwards by a few degrees and the reductions in the south are somewhat stronger (up to 30%). The rainfall reductions in winter are
The Mediterranean Climate Change under Global Warming
407
mainly due to the more frequent anticyclonic circulation in this region, which is imposed by the driving GCM (Giorgi et al., 2004). A strong decrease in precipitation is found for summer in all regions (widely more than 50%), again with different foci dependent on the driving models. For ECHAM4, the maximum reduction is over France but for the HadCM3 model, over the central Mediterranean Sea (Giorgi et al., 2004; Ra¨isa¨nen et al., 2004). Over land, the reduced rainfall leads to increasing temperature maxima due to drying of the soil. Simulations carried out by Arribas et al. (2003) suggest that land degradation (which is not included in the simulations mentioned before) may even enforce the simulated increase in temperatures and decrease in rainfall. Giorgi et al. (2004) also investigated the future changes in the characteristics of daily rainfall events over the Mediterranean region. They identified a general decrease in the number of rainy days. This effect is accompained by an increase in the intensity of daily rainfall in spring and autumn, while there is a decrease in the intensity of daily rainfall events mainly in summer and winter. These changes in the characteristics of daily rainfall can be accompanied by changes in the intensity of heavy precipitation events. Frei et al. (1998) found, for instance, a 25% increase in the frequency of heavy precipitation events exceeding 30 mm/day in southern Europe for the month of October due to the anticipated greenhouse warming. Giannakoupolous and Palutikof (2003) presented first results of a study dealing with the changing extremes in the Mediterranean region. On the basis of simulations until the year 2100 performed with the global low-resolution HadCM3 and the respective regional model HadRM3 they found increases in the number of tropical nights (from about 20 in a period with greenhouse gas forcing of the 1960s to about 100 in the year 2100) and, correspondingly, in the number of heatwaves over land. The increases were larger for the western Mediterranean area than for the central and eastern Mediterranean region.
8.3.3. AGCM Simulations with Variable Resolution The LMDZ-Mediterranean model is a variable-grid global AGCM, with a resolution of about 160 km over the Mediterranean region. The control simulation uses the present-day climate, represented by conditions at the end of the twentieth century (LMDZ/CTRL). Three simulations are performed for future scenario at the end of the twenty-first century. They differ by their boundary conditions coming from three global-coupled climate models from IPSL, CNRM and GFDL, respectively, all using the A2 emission scenario (Houghton et al., 2001). The three regional-zoomed simulations will be referred to as LMDZ/IPSL, LMDZ/CNRM and LMDZ/GFDL.
408 Mediterranean Climate Variability
Figure 135: Annual mean changes of the surface air temperature (K; left column) and precipitation (mm/day; right column) for the end of the twenty-first century, simulated by the LMDZ AGCM with stretched grids over the Mediterranean Sea. From top to bottom are three climate scenarios for the IPCC-A2 emission scenario as given by three different global coupled climate models, i.e. IPSL, CNRM and GFDL, respectively. The left panel of Fig. 135 displays the annual-mean changes in the surface air temperature for the three simulations. The spatial pattern is similar for the three runs, with a warming from 2 to 3 K over the sea, and 3–4 K for the surrounding lands. LMDZ/IPSL gives the smallest warming and LMDZ/GFDL the largest.
The Mediterranean Climate Change under Global Warming
409
The right panel of Fig. 135 gives the changes in precipitation. There is a general increase of precipitation for the North of Europe and a decrease for the South, including the Mediterranean basin. This general tendency is in agreement with other results reviewed here. Several centres of negative anomalies are located in northern Spain and southern France, over the land mass between the Adriatic Sea and the Aegean Sea, and in Turkey. The magnitude is rather small in LMDZ/GFDL (about 0.1 mm/day) and quite large in LMDZ/IPSL and LMDZ/CNRM (about 0.6 mm/day). Regional water cycle is an important element of the Mediterranean climate. As pointed out in Chapter 7 of this book, the Mediterranean Sea is a concentration basin with evaporation much larger than rainfall (Mariotti et al., 2002). Furthermore, the Mediterranean water cycle is also characterized by a strong seasonal cycle. Figure 136 displays the seasonal variations of the atmospheric part of the Mediterranean water cycle. i.e. precipitation (P), evaporation (E) and water deficit (E P), in the control simulation LMDZ/CTRL (upper panel) against, observation-based values (lower panel; Mariotti et al., 2002). Annualmean values are also given in the legend. The seasonal cycle of the simulated precipitation is quite realistic with a minimum in July–August and a maximum in November–December, although the absolute rainfall seems to be underestimated by the model. Evaporation, on the other hand, is overestimated by the model, and the seasonal cycle presents a one-month temporal shift: the minimum is reached in May for observation, but in June for the simulation. As a consequence, the water deficit budget is overestimated by the model and the one-month seasonal shift is also visible. Table 9 gives the future changes of the annual-mean values for changes in E, P and E minus P for the three simulations. All the three simulations show a decrease of precipitation rate, while evaporation increases for LMDZ/IPSL and LMDZ/CNRM and slightly decreases for LMDZ/GFDL. The net water deficit thus increases in the three scenarios, about 10% for LMDZ/IPSL, 14% for LMDZ/CNRM, but only 1.3% for LMDZ/GFDL. The last line of Table 9 shows the changes in the gain of the total heat flux at the sea surface for the three scenarios compared to the control simulation. Instead of losing heat into the atmosphere as in the control simulation ( 2.1 W/m2), the Mediterranean gains energy from the atmosphere for future scenarios. The net gain of heat flux varies from 3.6 to 11.9 W/m2 for different runs. Gibelin and De´que´ (2003) used the ARPEGE-CLIMATE global AGCM with variable resolution in order to study the anthropogenic climate change over the Mediterranean. Their model has a resolution of about 0.5 over the Mediterranean region and about 4.5 at the lowest resolution point close to New Zealand. The high-resolution atmospheric GCM is able to simulate many of the regional details that low-resolution coupled climate models cannot resolve
1600
E - P (970 mm/yr) E (1256 mm/yr) P (286 mm/yr)
E
1400 1200
E - P
1000 800 600 P 400 200 0
1
2
3
4
5
6
7
8
9
10
11
12
11
12
Mediterranean hydrological cycle (from Mariotti et al. 2002) 1800 E-P (627 mm/yr) E (1042 mm/yr) P (415 mm/yr)
1600 1400 1200
E
1000 E - P
800 600
P
400 200 0
1
2
3
4
5
6
7
8
9
10
Figure 136: Monthly mean changes of precipitation (P), evaporation (E) and the difference E P over the Mediterranean Sea for the present-day climate simulated by LMDZ/CTRL (upper panel) and observations (lower panel).
410 Mediterranean Climate Variability
Mediterranean hydrologic cycle (control simulation) 1800
The Mediterranean Climate Change under Global Warming
411
Table 9: Annual-mean changes of evaporation, precipitation, water deficit and gain of heat for the whole Mediterranean Sea and for the three scenarios respectively. LMDZ/IPSL Evaporation (E) (mm/year) Precipitation (P) (mm/year) E P (mm/year) Gain of heat flux (W/m2)
39 57 96 3.6
LMDZ/CNRM 57 74 131 5.8
LMDZ/GFDL 7 20 13 11.9
and, hence, able to add regional details to the prediction of the potential future climate change. Comparing the scenario period (2070–2099) and the present-day period (1960–1989), the simulation predicts a warming and drying of the Mediterranean region in the future. The decrease in mean precipitation is associated with a significant decrease in soil wetness and could have a considerable impact on the water resources around the Mediterranean basin. Somot (2005) has investigated the relation of the precipitation changes in this model with respect to its association with cyclones. In agreement with the studies mentioned in Section 8.2, a statistically significant 16% decrease in the number of intense cyclones was found (minimum intensity of 1.5 10 4 s 1) which is more emphasized in winter. Regionally, the Gulf of Genoa and the Aegean/Turkey area are most affected. For the Gulf of Genoa area, the simulated precipitation associated with the cyclones was studied by the mean of composites for the time of the maximum 850 hPa vorticity, which must exceed a minimum strength of 1.5 10 4 s 1. In summer, the amount of precipitation associated with a typical cyclone decreases up to 30% going from 18.5 to 13.0 mm/day (average in a 50-km radius circle around the cyclone centre). This result is similar for cyclones born in the Aegean Sea/Turkey Mountains area. This large decrease could explain at least in part the drying observed in the IPCC A2 scenario (Gibelin and De´que´, 2003) for the Mediterranean region in summer. Contrary to the summer situation, the cyclones associated precipitation increases in autumn (þ17%, 21 mm/day in the scenario) and spring (þ23%, 13 mm/day in the scenario) for the same Northern Mediterranean areas. In the present-day climate, autumn and spring are already the seasons during which intense rainy events occur around the Mediterranean basin, often associated with cyclones. Consequently, this kind of events might increase at the end of the twenty-first century following our simulation. Finally, no change is observed for winter precipitation composites with a typical value of 11.0 mm/day in the scenario as in the present-day climate.
412 Mediterranean Climate Variability
8.4. Impacts on Land Surface and Vegetation The climate system does not consist only of the atmosphere and ocean. For the Mediterranean area, the biosphere and the soils are parts of the system that have only recently been taken into account. In addition to their role as part of the system, they are important with respect to impacts of climatic change. Only a few studies are available which elucidate potential future tendencies. Taking tree ring growth under CO2 doubling as a parameter to quantify climate change impacts on the biosphere, Keller et al. (2000) considered regional effects on the French parts of the Mediterranean. They used meteorological data from the Arpe`ge-CLIMATE model, simulating a temperature increase of 3 K and a small increase of precipitation. Results for 5 tree species were obtained using an approach which involved an artificial neural network and an empirical tree ring model. A sensitivity to climate change was only found for few populations located at the boundaries of their ecological area, including enhanced growth of high altitude populations as well as growth reduction due to water stress in summer. As soil wetness is a parameter used in atmospheric models, it has been considered more frequently, but with the underlying assumption that there is no change in the (parameterized) vegetation. A common result (Wetherald and Manabe, 2002; Manabe et al., 2004) is reductions in soil moisture on the Mediterranean coast of Europe, similar to any other semiarid regions of the world. The relative change is particularly large during the dry season. The role of the land surface and of vegetation under climate change can be expected to be more complex than what is found out by simulating it under the precondition of constant vegetation types, soils or even the complete absence of feedbacks. Even direct anthropogenic effects on vegetation must be taken into account. A study going into parts of these problems was performed by Du¨menil-Gates and Liess (2001). They estimated the impacts of deforestation in the Mediterranean region as they are arising from the particular parameterizations used in a GCM. They found a reduction of precipitation in summer, resulting from lower plant evapotranspiration and lower evaporation from soils, assuming that they are eroded. The occurrence and the impacts of erosion under climate change have been explored for a small forested catchment in Catalonia. Given a 4-K temperature increase, Avila et al. (1996) have investigated the effects of a simultaneous 10% increase or decrease in precipitation. They found enhanced weathering rates in the case when only rainfall was increased, and that the resulting effects on runoff water chemistry were much larger than those arising from changing atmospheric input of Sahelo-Saharan dust. The small number and diversity of research results make it clear that more comprehensive work must be performed into the complex
The Mediterranean Climate Change under Global Warming
413
reaction of the climate system to the imposed rise of greenhouse gas and other anthropogenic forcing.
8.5. Future Research It appears that the number of studies on the future climate in the Mediterranean region is still rather sparse. Future research needs to comprise studies evaluating the stability of results produced by the different global and regional climate models. It is still an open question whether the differences between simulated changes are due to specific characteristics of the models, in particular due to the parameterizations in the models, or whether they are due to internal decadal variability in the climate system. The coupling of atmospheric models with models of the Mediterranean Sea (Chapter 7) is also an important aspect for the progress in the assessment of climatic change in this region. Aspects that clearly need more attention are the occurrence of different kinds of extreme weather events or the complex feedbacks with the biosphere and soils. The combinations of climatic features that produce the particular vulnerability of the Mediterranean environment (e.g. irregular or scarce water availability combined with an occurence of floods or heat waves) should be considered in depth. Future research should include both efforts for a better understanding of the mechanisms of climate changes and statistical approaches. Multi-model ensemble simulations with both global and regional models help to assess the uncertainties of climatic change and to provide probabilistic estimates of long-term changes. In order to minimise the risks of climate change, it is also important to perform more research into the adaptation of its effects. This requires more joint work of many different scientific disciplines. In this context, it is a goal of future research to establish earth models that integrate all aspects of the climate system simultaneously. As this will require several decades of research work it is presently important to continue work employing more specialized approaches. Their subsequent integration into joint interdisciplinary research projects on particular aspects of the (regional) earth system under climate change could be a promising perspective, in particular for the Mediterranean basin.
References Arribas, A., Gallardo, C., Gaertner, M. A., & Castro, M. (2003). Sensitivity of the Iberian Peninsula climate to a land degradation. Climate Dynamics, 20, 477–489. Avila, A., Neal, C., & Terradas, J. (1996). Climate change implications for streamflow and streamwater chemistry in a Mediterranean catchment. Journal of Hydrology, 177, 99–116.
414 Mediterranean Climate Variability
De´que´, M., Marquet, P., & Jones, R. G. (1998). Simulation of climate change over Europe using a global variable resolution general circulation model. Climate Dynamics, 14, 173–189. Du¨menil-Gates, L., & Liess, S. (2001). Impacts of deforestation and afforestation in the Mediterranean region as simulated by the MPI atmospheric GCM. Global and Planetary Change, 30, 309–328. Frei, C., Scha¨r, C., Lu¨thi, D., & Davies, H. C. (1998). Heavy precipitation processes in a warmer climate. Geophysical Research Letters, 25, 1431–1434. Giannakoupolos, C., & Palutikof, J. P. (2003). Modeling climate extreme events: the EU project MICE for the case study area of Greece. The Eggs, 5, 29–35, available on-line at www.the-eggs.org. Gibelin, A.-L., & De´que´, M. (2003). Anthropogenic climate change over the Mediterranean region simulated by a global variable resolution model. Climate Dynamics, 20, 327–339. Giorgi, F., & Francisco, R. (2000). Evaluating uncertainties in the prediction of regional climate change. Geophysical Research Letters, 27, 1295–1298. Giorgi, F., Hewitson, B., Christensen, J. H., Hulme, M., vonStorch, H., Whetton, P., Jones, R., Mearns, L. O., & Fu, C. (2001a). Chapter 10: Regional climate information – evaluation and projections. In: J. T. Houghton, Y. Ding, D. J. Griggs, M. Noguer, P. J. van der Linden, & D. Xiaoxu (Eds)., Climate Change 2001: The Scientific Basis, Contribution of Working Group I to the Third Assessment Report of the Intergovernmental Panel on Climate Change, (Cambridge University Press, Cambridge, UK, pp. 583–638). Giorgi, F., Whetton, P. W., Jones, R. G., Christensen, J. H., Mearns, L. O., Hewitson, B., vonStorch, H., Francisco, R., & Jack, C. (2001b). Emerging patterns of simulated regional climatic changes for the 21st century due to anthropogenic forcings. Geophysical Research Letters, 28, 3317–3320. Giorgi, F., Bi, X., & Pal, J. (2004). Mean, interannual variability and trends in a regional climate change experiment over Europe. II: Climate change scenarios (2071–2100). Climate Dynamics, 23, 839–858. Gonza´lez-Rouco, J. F., Heyen, H., Zorita, E., & Valero, F. (2000). Agreement between observed rainfall trends and climate change simulations in the southwest of Europe. J. Climate, 13, 3057–3065. Houghton, J. T., Ding, Y., Griggs, D. J., Noguer, M., van der Linden, P. J., Dai, X., Maskell, K., & Johnson, C. A. (Eds). (2001). Climate change 2001: the scientific basis. Cambridge University Press, Cambridge, UK, 881 pp. Keller, T., Edouard, J. L., Guibal, F., Guiot, L., Tessier, L., & Vila, B. (2000). Impacts of a climate warming scenario on tree growth. (Comptes Rendus de l’Academie des Sciences, Serie III, Sciences de la vie), 323, pp. 913–924. Kittel, T. G. F., Giorgi, F., & Meehl, G. A. (1998). Intercomparison of regional biases and doubled CO2 sensitivity of coupled atmosphere-ocean general circulation model experiments. Climate Dynamics, 14, 1–15. Knippertz, P., Christoph, M., & Speth, P. (2003). Long-term precipitation variability in Morocco and the link to the large-scale circulation in recent and future climates. Meteorology and Atmospheric Physics, 83, 67–88. Leckebusch, G. C., Koffi, B., Ulbrich, U., Zacharias, S., Pinto, J. G., & Spangehl, T. (2005). Analysis of frequency and intensity of winter storm events in Europe on synoptic and regional scales from a multi-model perspective. Climate Research, in review.
The Mediterranean Climate Change under Global Warming
415
Lionello, P., Dalan, F., & Elvini, E. (2002). Cyclones in the Mediterranean region: the present and the doubled CO2 climate scenarios. Climate Research, 22, 147–159. Lionello, P., Elvini, E., & Nizzero, A. (2003). Ocean waves and storm surges in the Adriatic Sea: intercomparison between the present and the doubled CO2 climate scenarios. Climate Research, 23, 217–231. Lionello, P., Cogo, S., Galati, M. B., & Sanna, A. (2006). Climate scenario simulations of the wind wave field in the Adriatic Sea (in preparation). Manabe, S., Milly, P. C. D., & Wetherald, R. (2004). Simulated long-term changes in river discharge and soil moisture due to global warming. Hydrological Sciences Journal, 49, 625–642. Mariotti, A., Struglia, M. V., Zeng, N., & Lau, K.-M. (2002). The hydrological cycle in the Mediterranean region and implications for the water budget in the Mediterranean Sea. Journal of Climate, 15, 1674–1690. Parry, M. L. (Ed). (2000). Assessment of potential effects and adaptations for climate change in Europe. The Europe ACACIA project. Jackson Environment Institute, University of Norwich, UK, 320 pp. Pinto, J. G., Spangehl, T., Ulbrich, U., & Speth, P. (2005a). Sensitivities of a cyclone detection and tracking algorithm: individual tracks and climatology. Meteorologische Zeitschrift, submitted. Pinto, J. G., Spangehl, T., Ulbrich, U., & Speth, P. (2005b). Assessment of winter cyclone activity in a transient ECHAM4-OPYC3 GHG experiment. Meteorologische Zeitschrift, 14, 823–838. Ra¨isa¨nen, J., Hansson, U., Ullerstig, A., Doescher, R., Graham, L. P., Jones, C., Meier, H. E. M., Samuelson, P., & Willen, U. (2004). European climate in the late twenty-first century: regional simulations with two driving climate models and two scenarios. Climate Dynamics, 22, 13–31. Sanchez, E., Gallardo, C., Gaertner, M. A., Arribas, A., & Castro, M. (2004). Future climate extreme events in the mediterranean simulated by a regional climate model: a first approach. Global and Planetary Change, 44, 163–180. Somot, S. (2005). Modelisation climatique du bassin Mediterranean: variabilite et sce´narios de changement climatique. PhD Thesis. Universite´ Paul Sabatier, Toulouse, France, 333 pp. (in French). Sumner, G. N., Romero, R., Homar, V., Ramis, C., Alonso, S., & Zorita, E. (2003). An estimate of the effects of climate change on the rainfall in Mediterranean Spain by the late twenty first century. Climate Dynamics, 20, 789–805. Trigo, R. M., & Palutikof, J. P. (2001). Precipitation scenarios over Iberia: a comparison between direct GCM output and different downscaling techniques. Journal of Climate, 14, 4422–4446. Ulbrich, U., & Christoph, M. (1999). A shift of the NAO and increasing storm track activity over Europe due to anthropogenic greenhouse gas forcing. Climate Dynamics, 15, 551–559. Ulbrich, U., Christoph, M., Pinto, J. G., & Corte-Real, J. (1999). Dependence of winter precipitation over Portugal on NAO and baroclinic wave activity. International Journal of Climatology, 19, 379–390. Wetherald, R. T., & Manabe, S. (2002). Simulation of hydrologic changes associated with global warming. Journal of Geophysical Research, 107, 10029/2001JD001195.
Subject Index ACACIA (A Consortium for the Application of Climate Impact Assessment), 400, 401 Adriatic Deep Water (ADW), 244, 259, 292, 296 Atlantic Water (AW), 6, 257, 285, 314, 375 Aegean Sea, 296, 297, 315 aerosols, 400 air–sea interaction, 229 All-India Rainfall Index (ARI), 160 analyses, 228, 332, 334–336, 343 dataset, 327, 332, 333 objective, 328, 332, 335, 344 subjective, 328, 332–334, 344 anthropogenic climate change, 400, 409 anthropogenic global warming, 377 anticyclone, 160 Azores, 213 Arctic Oscillation (AO), 69 arrival times, 287, 290, 291, 293 Atlantic inflow, 228, 235, 239, 260 atmosphere–ocean general circulation model (AOGCM), 113, 114, 115 atmospheric circulation, 32, 36, 62, 64, 98, 105, 113, 121, 124 atmospheric general circulation model (AGCM), 400, 407, 411 high resolution, 400, 409 physical parameterization, 398, 410 variable resolution, 404, 407, 409 Azores High, 373, 375 Black Sea, 236–239 outflow, 237, 246, 256–259
blocking, 189–192, 201, 205, 339, 359 pattern, 99 episode, 182, 190, 191 condition, 201 canonical correlation analysis (CCA), 182, 192, 193, 197 carbon dioxide (CO2), 400–402, 404, 412 circulation, 406 climate impact, 183, 192, 204, 213, 216 scenario, 217 trend, 180, 183, 198–212, 216 variability, 183, 191, 210–212, 218 extreme, 29, 31, 38, 92, 97, 120 pattern, 260 climate field reconstruction (CFR), 77, 78 cold intermediate layer, 236 composite analysis, 83–91 concentration basin, 338 convection, 293, 298–300, 307, 316, 391 coral, 30, 57, 60, 69–71, 73–79, 120, 121 coupled climate model, 404, 405, 409 ocean–atmosphere model, 113 cyclogenesis, 328–331, 333, 337–339, 341, 342, 347, 353, 362, 363 areas, 328, 329, 331, 333, 334, 347, 351, 359, 363 mechanisms, 327, 330, 337, 338, 341, 344, 353, 354, 362, 364 orographic, 330, 341, 342, 346, 351
418 Subject Index
cyclone, 205, 325–338, 340–347, 349–351, 353–355, 357, 359–365, 400–402, 404, 411 Atlantic, 180, 185, 205, 341, 342, 347 hurricane-like, 330, 331, 342, 362, 364 Lee, 330, 331, 362, 363 mid-latitude, 180 northern African, 331, 338, 341 seasonality, 327, 331, 337, 338, 362, 363 thermal low, 330, 331, 362, 363 tropical, 327, 331 deep water formation, 228–230, 232, 238, 240, 241, 243, 245, 246, 252, 253, 255, 256, 259, 261, 268, 270 Aegean, 228, 230, 232, 237, 238, 242–245, 251, 255–262, 264 Adriatic, 228, 230, 234, 240, 242, 244, 245, 251, 257, 259, 260, 262, 264 Western Mediterranean, 228, 240–243, 245–247, 260, 262, 263 deep water trend, 243–247 Black Sea, 246 Eastern Mediterranean, 243–245 Western Mediterranean, 245, 246 deforestation, 412 documentary data, 30, 32, 38, 46, 51, 53, 78, 79, 83, 119 proxy, 32, 33, 39, 45, 52, 78, 120, 122 downscaling, 399, 400, 404, 406 dynamical, 400, 404 statistical, 399, 405 drought, 33, 34, 37, 40, 50, 58, 59, 62, 64, 65, 98, 112, 113 Eastern Mediterranean Deep Water (EMDW), 243–245 Eastern Mediterranean Transient (EMT), 230, 249, 251, 260, 263
evaporation–precipitation (E–P), 231, 232, 235, 259, 261, 262 ECMWF (European Centre for Medium range Weather Forecast), 374, 379–382 economic damages, 326, 361, 363 El Nin˜o Southern Oscillation (ENSO), 32, 109, 113, 116, 149, 150, 212 impact, 212 El Nin˜o, 32, 149, 151, 152–154, 156, 157, 159, 212, 213 La Nin˜a, 150–154, 157, 212 empirical orthogonal function (EOF), 197, 294, 295 EMULATE, 332, 344 entrainment, 300, 311 ERA (ECMWF ReAnalysis), 379, 381 evaporation, 228, 231, 238, 239, 241, 242, 246, 259, 261, 405–408 external forcing, 32, 78, 114, 115, 117, 118, 123, 124 extreme, 326, 356, 359, 363–365 floods, 325–327, 333, 342, 346, 347, 350, 355 feedback, feedbacks, 10, 157, 276, 286, 293, 412, 413 mechanism, 11, 162, 293 convective, 307 advective, 307–309 air–sea, 10 climate, 392 general circulation model (GCM), 113, 115, 118, 119, 217 geopotential high, 187, 190, 192, 193, 200, 212 greenhouse gas (GHG), 31, 115 scenario, 405 Gulf of Lions, 286, 293, 297, 299, 316 Hadley cell, 373, 381 Hadley Climate Model (HadCM3), 32, 113–116, 122, 125
Subject Index 419
heat, 4, 45, 186, 245, 258, 285, 305, 330, 375, 388, 390, 411 loss, 10, 230, 232, 244, 245, 256, 261, 383, 388 fluxes, 8, 189, 213, 230, 257, 296, 349, 379, 388, 390, 409 heatwave, 407 hurricane, 149, 162 hydraulic control, 300, 316 Icelandic low, 373, 375, 377 impact, 375 Intermediate Water (IW), 292 IPCC (Intergovernmental Panel for Climate Change), 381 jet-stream, 377, 378 sub-polar, 378 sub-tropical, 375 Jordan river discharge, 151 Lagrangian diagnostics, 287–289 anomaly propagation, 283, 292, 303, 306, 309 circulation path, 243 hydrological properties, 292, 293 propagation time scale, 292 lake sediments, 30, 59 landslides, 325, 327, 328, 344, 360, 361, 363 Levantine Intermediate Water (LIW), 228, 242 Little Ice Age (LIA), 45, 52, 63, 68, 69, 74, 75, 97 meddies, 305, 306, 310 MEDEX (MEDiterranean EXperiment), 326, 332, 364 medieval warm period (MWP), 52, 68, 69, 74 Mediterranean, 29, 30–33, 39, 45, 49, 51–61, 63 climate, 149, 154, 156, 158–160, 168, 170, 171
lows, 331, 341, 364 oscillation, 190, 202, 211 outflow, 282, 284, 300, 301, 303, 305, 306–311 salinity tongue, 312, 314 thermohaline circulation, 285–287, 289, 315 undercurrent, 305, 306, 312 Mediterranean Outflow Water (MOW), 283, 284, 286, 292, 293, 303–312, 314, 316, 376, 383, 384, 391 mixed layer, 379, 380 model–data intercomparison, 113–119 model, 376–381, 382–385, 387, 388, 390, 391 Arpege, 381, 382, 387, 388 ECHO-G, 114 ensemble, 413 LMDZ, 377, 378 MED16, 379 MED8, 379–381, 382, 387, 388 OPA, 379, 387 SAMM, 387, 388 variable grid, 386 Modified Atlantic Water (MAW), 239–241 monsoon, 159–162, 266, 267 African, 149, 160–162, 171 South Asian, 149, 159, 161, 170 waves, 266–267 multiple equilibria, 296 multi-proxy reconstruction, 124 natural disaster, 31, 38 NCEP, 332, 343, 344, 364 near-surface air temperature, 84, 85, 90, 109–111, 121, 123 North Atlantic, 283, 284, 286, 289, 301, 303, 312, 314, 316 North Atlantic Oscillation (NAO), 41, 43, 58, 67–70, 78, 88, 90, 99, 105–109, 115, 116–119, 121, 154, 156–158, 161, 170, 182–187, 189, 191, 195, 197, 204, 205, 207, 210–213, 215–217
420 Subject Index
Northern Adriatic Dense Water (NADW), 244 Northern Hemisphere (NH), 31, 107 numerical model, 285, 293, 300 ocean waves, 327, 357–362 trends, 327, 328, 332, 333, 343, 344, 349, 350, 356, 359, 362 orography, 2, 6, 164, 180, 329, 330, 335, 338, 341, 349, 351, 362, 363 overturning cell, 286, 295 paleoclimate, 393 paleoclimatology, 44, 79 Palmer Drought Severity Index (PDSI), 32, 95, 96 past climate, 29, 31, 32, 44, 48, 51, 71 variability, 29, 31, 69, 71, 83, 117, 123 pattern East Atlantic (EA), 183, 187, 189, 191 East Atlantic Jet (EA-Jet), 191, 192 East Atlantic/Western Russia (EA/WRUS), 106, 121, 157, 187, 205, 216, 342, 363 Scandinavian (SCAND), 183, 188–190 Southern Europe Northern Atlantic, 342 synoptic, 353 POEM, 5 precipitation, 31–35, 38, 40–45, 48, 50, 52, 58–60, 62–70, 77–85, 87–95, 97, 99, 103, 105, 106, 108, 109, 113, 116, 117, 120–124, 171, 180, 195–198, 202–212, 215–218, 231, 232, 238, 239, 247, 260, 261, 325–328, 337, 343–347, 349, 350, 352, 363–365, 401, 402, 405–412 extremes, 180 variability, 193, 195, 196 trend, 202, 205, 215, 216 pressure lows, 351, 353, 354, 363
principal component analysis (PCA), 99, 194 principal components (PCs), 105, 106, 107 proxy, 30–34, 36, 38, 41, 50, 53, 57, 59, 66, 68, 73, 77, 106, 113, 151, 307 PRUDENCE, 406 Radiative Forcing (RF), 114, 115, 118 rain, 326, 344–348, 350, 363, 365 rapid changes, 196, 203, 205, 332 reanalysis, 196, 203, 205, 332 Red Sea Trough, 149, 164, 165, 171, 172 regional climate model, 400, 413 resource management, 120 rivers, 229, 232, 233, 239, 246, 257, 259, 269 Black Sea, 239 Mediterranean, 233 Rossby wave, 151, 161 running correlation analysis, 105, 121 Sahara dust, 149, 168 salinity, 10, 14, 73, 76, 123, 228, 235, 237, 241, 242, 244–248, 250, 253–255, 283, 285, 296, 300, 301, 303, 309, 310–312, 314–316, 379, 383, 384 sea level, 238, 260–263 Black Sea, 236, 238, 261 extremes, 263–265 Mediterranean, 261–263 trends, 262, 263 sea level pressure (SLP), 66, 111, 159, 184, 185, 192, 197, 198, 210, 211 sea surface temperature (SST), 69, 73, 75, 76, 109, 113, 123, 192–194, 200, 213, 262 seasonal variability, 236, 238 ship logbooks, 38, 53, 120, 122 Sicily channel, 286, 288, 292, 299, 314, 316 simulation, 379, 384, 386, 388–391 ensemble, 377, 385 equilibrium, 377 transient, 377
Subject Index 421
socio-economic impacts, 29, 120 soil moisture, 408 solar variability, 213, 215 solar volcanic radiative forcing (SVRF), 118 speleothem, 29, 30, 57, 60, 67, 68, 74, 76, 79, 120, 121 storm surge, 325, 333, 353–357, 363 storm track, 180, 186, 189, 205, 218, 337 straits, 235, 238, 251 Bosphorus, 238 Cretan, 251, 257 Dardanelles, 238 Gibraltar, 228, 229, 235, 237, 241, 242, 263 Otranto, 242, 244 Sicily, 240–242, 257 subtropical jet, 164, 165 sunspots, 357 surge level, 404 synoptic variability, 210, 212, 333 Syrian lows, 331, 347 teleconnection, 32, 81, 109, 112, 113, 122, 123, 149, 150, 154, 158, 161, 168, 170, 171, 376, 377 temperature, 180–183, 186, 189, 190, 192–194, 198–202, 211–213, 216–218, 406–408, 412 extremes, 407 trend, 198–202 variabiltity, 192, 193 thermohaline circulation, 379, 381, 383, 387 global cell, 288, 289 lower branch, 284, 286, 289, 290, 292, 293
upper branch, 285–287 western cell, 288, 289 thermal lows, 330, 331, 362, 363 TOPEX/POSEIDON, 262 transient eddies, 157 tree-ring, 57, 63, 64, 66, 67, 78, 79, 121, 123 tropical Atlantic variability, 161 tropical night, 407 water budget, 238 water mass, 292, 293, 297, 298, 300, 303, 306, 309, 314, 315, 375, 400 transformation, 287, 303 Western Mediterranean Deep Water (WMDW), 245–247, 260, 300, 375, 379, 383, 384 water resources, 411 wave height, 404, 405 waves, 264, 267, 325–328, 333, 344, 357–359, 362, 363, 365 wind stress, 379, 381, 383 wind wave, 14, 229, 265, 328, 357, 365 winds, 238, 244, 264, 326, 327, 333, 344, 351, 353, 354, 357, 359, 362, 365 Bora, 351, 359 ECMWF, 258, 259, 265 Etesian, 351, 353, 359 Mistral, 351, 354, 359 Sirocco, 351, 354, 359, 363 Vandevales, 134 Zonal Overturning Stream Function (ZOF), 384, 385
This page intentionally left blank